Key West Oct 17 - 21 Oct 20 - 24

2018

Abstract Booklet 10th Intl. Conference on Charged Particle Optics

13th Intl. Computational Accelerator Physics Conference

Platinum Sponsors of CPO-10

Silver Sponsors of CPO-10

Gold Sponsors of CPO-10

Platinum Sponsor of ICAP’18

Silver Sponsors of ICAP’18

CPO-10 ABSTRACTS

A HIGH BRIGHTNESS FIELD EMITTER BY USEOF NOBLE METAL COATED NANO SCALE

PYRAMID FORMED ON TUNGSTEN TIP

Hirotaka Asai, Ryota Kawai, Fumiya Matsubara,Hidekazu Murata, and Eiji Rokuta

Faculty of Science and Engineering, Meijo University

Keywords: Nano tip; field ion microscopy; field emissionmicroscopy

Nano tips have attracted large attention as the novel electronsource with high brightness in high performance electron micro-scopes. In field emission electron guns, the brightness is im-proved as the width of electron source is decreased. This is be-cause the electric field converges to the single atom at the tip end,resulting in a dramatic collimated electron beams. Tungsten (W)nanopyramids coated with monolayer films of noble metal (NM)are potential for a material of highly coherent electron sources.Sides of a tungsten (W) needle was coated with collodion con-taining powder of palladium oxide (PdO) at first. The W nee-dle was subsequently annealed under ultrahigh vacuum to sup-ply palladium (Pd) atoms to an apex of the W needle via surfacediffusion and to produce a W nanotip coated with monolayer Pdfilms. Field ion microscopy (FIM) revealed {111} planes of theW nanotip were contracted by faceting 211 planes surroundingthe 111 plane, of which the structural change is analogous to aformation of three sided W nanopyramids. General appearancesof the W nanopyramids were the same as those shown by the ex-isting naopyramids. The opening angle corresponding to FWHMof the beam profile of a nanopyramid with a top consisting of fif-teen atoms was about 9.4 degree at extractor voltage of 711 Vand emission current of 0.6 nA. In order to evaluate the bright-ness of the single atom tip, we adopted a regular triangle with aside of 0.18 nm was regarded as emission region and source areaestimated 1.4E-14 cm2. We evaluated the brightness of 2.0E4 Astr−1 cm−2 V−1. The evaluated value is higher than that of theconventional FE source (5.0E3 A str−1 cm−2 V−1).

——————————————–

THE ION-OPTICAL DESIGN OF THE HIGHRIGIDITY SPECTROMETER HRS FOR FRIB

G.P.A.Berg

University of Notre dame

Invited TalkKeywords: Spectrometer, Ion-optical design, Rare Isotopes

With the ongoing construction of the Facility for Rare-IsotopeBeams FRIB for the production of high-intensity rare isotopes(RI), several electro-magnetic experimental analysis systems areunder construction and design to exploit these beams for a wideexperimental science program. A large part of the science pro-gram can be executed using the HRS as outlined in the White Pa-per “HRS A High Rigidity Spectrometer for FRIB” (December

2014, ed. A. Gade and R. Zegers). The HRS will enable gainsin luminosity for experiments with rare-isotope beams at FRIBby factors of 2-100, with the highest gains for the most neutron-rich isotopes, including those in the path of the astrophysical r-process. Therefore, the HRS will add tremendously to the dis-covery potential of FRIB. In this presentation the ion-optical de-sign of a beam line/spectrometer system will be presented thatsatisfies all essential design requirements. This includes twomodes of operation for missing mass experiments with MoNA-LISA and high-resolution spectroscopy. Both modes can oper-ate with the Gamma Ray detector GRETA at the target and al-low full dispersion matching of beam line and spectrometer forhigh resolution without dramatically reducing the beam intensityusing momentum slits. The matching conditions also allow thereconstruction of the properties of the reaction products usingthe detector system. This work is supported by the U.S. Depart-ment of Energy Office of Science under Grant DE-SC0014554,ION-OPTICAL AND ASSOCIATED MAGNET FEASIBILITYSTUDY OF A HIGH RIGIDITY SPECTROMETER.

——————————————–

LENS-MIRROR OBJECTIVE FORTRANSMISSION ELECTRON MICROSCOPE

S.B. Bimurzaev, N.U. Aldiyarov, E.M. Yakushev

Almaty University of Power Engineering andTelecommunication

Keywords: lens-mirror objective, transmission electronmicroscope, spherical aberration, axial chromatic aberration

A new electron-optical scheme of the lens-mirror objective fortransmission electron microscopes (TEM), based on the specialfocusing regime (the so-called superimposed image mode) in thecenter of curvature of the deflecting magnetic field [1-3], is con-sidered. The magnetic field does not cause an additional distor-tion of the image, and the problem of calculating the objectiveis reduced to calculating a relatively simple lens-mirror systemwith a common rotational symmetry axis. The new data on theparameters of specific lens-mirror systems composed of a well-known magnetic lens with a bell-shaped distribution of the axialfield and an electrostatic mirror with electrodes in the form of aset of coaxial cylinders of equal diameter have been obtained. Arather wide family of mirror-lens electron-optical systems witha simultaneous compensation of the main types of aberrations(spherical and axial chromatic) with a large linear magnifica-tion has been found. The diffraction limit of the linear resolu-tion of the lens-mirror objective has been evaluated under thejoint action of the remaining fifth-order spherical aberrations andthe diffraction of electrons by the beam-limiting diaphragm. Ithas been shown that full elimination of the third-order spher-ical aberrations can significantly increase the resolution of thetransmission electron microscopes (TEM) and, even at moder-ate accelerating voltages about 100 kV, give high resolution val-ues of less than one Angstrom, inaccessible for modern high-voltage TEM / STEM devices. 1. Bimurzaev S.B. and Yaku-

1

shev E.M. (2013) Electron lens aberration corrector. // WIPOPatent Application WO/2013/077715 A1. 2. Yakushev EM.“Theory and Computation of Electron Mirrors: The Central Par-ticle Method.” In Advances in Imaging and Electron Physics, Ed:Hawkes PW. Elsevier. 2013. V.178, P.147-247. 3. BimurzaevS.B. Aldiyarov N.U. and Yakushev E.M. The objective lens ofthe electron microscope with correction of spherical and axialchromatic aberrations//Microscopy.-2017. – Vol. 66, Issue 5.- P.356–365.

——————————————–

PLANAR MULTIREFLECTIVE TIME-OF-FLIGHTMASS SPECTROMETER OF A SIMPLE DESIGN

S.B. Bimurzaev

Almaty University of Power Engineering andTelecommunication, Almaty, Kazakhstan

Keywords: mass spectrometer, time-of-flight focusing,time-of-flight dispersion, electrostatic mirror

A scheme of an improved version of a planar multi-reflectivetime-of-flight mass spectrometer [1], based on the use of two-dimensional electrostatic mirrors and an ion source with an ionaccelerator forming an inhomogeneous electrostatic field [2],which provide special focusing modes for ionic fluxes, is pro-posed. Numerical calculations have been used to determine theconditions that allow, along with the formation of a parallel ionflux, the time-of-flight focusing of ions in energy up to fourthorder in the four-electrode ion accelerator in which the first elec-trode has a planar shape and can be combined with the exit win-dow of the ionization region. The data that determine the condi-tions of time-of-flight focusing of ions in energy up to the thirdorder inclusive in a three-electrode two-dimensional electrostaticmirror in the regime of a plane mirror have been obtained [3].Mirror electrodes are pairs of parallel plates symmetrically lo-cated relative to the symmetry plane of the mirror field. Twovariants of plane mirrors are considered: 1) when the forwardand reverse branches of the trajectory coincide; 2) when the for-ward and reverse branches of the trajectory are symmetric withrespect to the plane of symmetry. It is shown that the time-of-flight dispersion of the mirror by mass in the second variant isseveral times higher than in the first variant. The use of an ionsource forming a parallel ion flux in combination with highlydispersive mirrors serves as the basis for simultaneous enhance-ment of resolution and sensitivity of the mass spectrometer. Ref-erences 1. Nazarenko L.M., Sekunova L.M. and Yakushev E.M.SU Patent 1725289 A1 (1992). 2. Bimurzaev S.B., YakushevE.M., Nazarenko L.M. Innovative package of the RK. Author’scertificate No. 87127 (2015). 3. Kelman V.M., Fedulina L.V.,Yakushev E.M. Deviation of parallel beams of charged parts bya plane electrostatic mirror // Zhurnal Tekhnicheskoi Fiziki, 41(1971). P. 1825-1831.

——————————————–

LONGITUDINAL BEAM DYNAMICS STUDIES ATTHE PIP-II INJECTOR TEST FACILITY

J.-P. Carneiro, B. Hanna, L. Prost, A. Saini, A.Shemyakin, D. Sun

Fermi National Accelerator Laboratory

Keywords: Fast Faraday Cup, Bunch Length, LongitudinalEmittance

The Proton Improvement Plan, Stage Two (PIP-II) is a programof upgrades proposed for the Fermilab injection complex, whichcentral part is an 800- MeV, 2-mA CW-compatible SRF linac. Aprototype of the PIP-II linac front end called PIP-II Injector Test(PIP2IT) is being built at Fermilab. As of now, a 15-mA DC,30-keV H- ion source, a 2 m-long Low Energy Beam Transport(LEBT), a 2.1-MeV CW RFQ, followed by a 10-m Medium En-ergy Beam Transport (MEBT) have been assembled and tested.A Fast Faraday Cup (FFC) installed in the MEBT measures thelength of a beamlet cut out of the bunch by a small-size entrancehole of the FFC. The information about the bunch length mea-sured at various settings allows for reconstruction of the longitu-dinal beam dynamics and optimization of injection into the firstcryomodule. These measurements are compared with simula-tions by the beam dynamics codes TRACEWIN and TRACK.The paper describes the experimental procedures of the bunchlength measurements with the FFC, presents the measurementresults, and compares them with simulations. One of importantexperimental observations, confirmed by simulations, is the de-pendence of the FFC beamlet length on the radial position acrossthe beam.

——————————————–

THE ORDER 4 ALGORITHM FORCYLINDRICALLY SYMMETRIC

ELECTROSTATICS

David Edwards Jr

IJL research center

Keywords: Finite Difference Method, FDM, order 4 algorithm,nine point algorithm

It has been ∼60 years since Emile Durand first reported thefourth order algorithm for cylindrically symmetric electrostatics(1957). It was immediately clear that using this algorithm thepotential could be calculated with precisions significantly higherthan the standard 5 point or order 2 algorithm. The solution cameat a price however and the price was that it was essentially impos-sible to implement for boundaries not lying on rows and columnsof meshpoints. And this situation is basically the same today asit was 60 years ago in spite of the many attempts at incorporat-ing this algorithm into electrostatics. However in 2014 a solutionfor the curved boundary problem was found for FDM. Unfor-tunately as the emphasis in that work was on the higher orderalgorithms and these were entirely too complex to be detailed noexplicit formulations could be given. This significantly limitedthe usefulness of that work. It has become clear however thatthe fourth order algorithm, which when implemented by the pro-cess for curved boundaries as described in the 2014 report wouldprovide significant gains in precision over the order 2 algorithm

2

and likely compete favorably with the current BEM and FEMformulations. It is thus to provide such a description of that im-plementation that the present work is directed. In particular itwill allow all points within the geometry, independent of theirproximity to the boundary, to use the same 4th order algorithm,the one of Durand. Also the construction employed will in factbe more direct than that used in the standard order 2 implementa-tion for curved boundaries at a cost of only a relatively insignifi-cant increase in computational time. In addition the treatment ofboundary singularities by a multi-region construction will be de-scribed as such singularities can negate any gains that the order4 algorithm might yield.

——————————————–

ADVANCED BEAM OPTIC OPTIONS FORBROOKHAVEN NATIONAL LABORATORY

ACCELERATOR TEST FACILITY BEAMLINE

Mikhail Fedurin

Brookhaven National Laboratory Accelerator Test Facility

Keywords: masked beam, ultrashort bunch, micro-bunching

The Accelerator Test Facility (ATF) at Brookhaven NationalLaboratory operates as a National User Facility supported bythe Accelerator Stewardship Program in the US DOE’s Office ofHigh Energy Physics. The facility presently provides high bright-ness 70 MeV electron beams and terawatt-class CO2 laser capa-bilities to support wide program in advanced accelerator R&D.Present design of ATF beamline transport with beam manipula-tion tools and beam diagnostics together with proposed schemesto generate ultrashort and flat electron bunches will be discussed.

——————————————–

GENERALIZATION OF PARAXIAL TRAJECTORYMETHOD FOR THE ANALYSIS OF

NON-PARAXIAL RAYS: ELECTRON GUNDESIGNS IN TERMS OF OPTICAL PARAMETERS

Shin Fujita

Shimadzu Asia Pacific Pte Ltd.

Invited TalkKeywords: ray tracing, numerical calculation, cathode lens

The paraxial trajectory method has been generalized for appli-cation to the cathode rays inside electron guns. The generalizedmethod can handle rays that initially make a large angle with theoptical axis. The key to success of the generalization is the adop-tion of the trigonometric function sine for the trajectory slopespecification, instead of the conventional use of the tangent. Animproved assignment of paraxial trajectory to the actual ray be-comes possible by the new slope specification. It is possible torelate the ray emittance condition (the combination of positionand slope of rays at reference planes) on the cathode to thoseat the crossover plane using polynomial functions, whose coef-ficients can be used as the optical parameters in electron sourcecharacterization. The most important among the parameters isthe Electron Gun Focal Length, which can be used for quantita-tive estimate of both the crossover size and the angular current

intensity. Electron gun simulation program G-optk has been de-veloped based on the generalized paraxial trajectory theory. Theprogram calculates the principal paraxial trajectories, optical pa-rameters, as well as virtual emittances solely from the axial po-tentials and fields. It gives a clear physical picture of electronsources and can be used for the gun design optimization.

——————————————–

ELUCIDATION OF ION MOTION INQUADRUPOLE MASS FILTER BY BLOCH

FUNCTION: IMPROVED PRE-ROD DESIGN FOREFFICIENT ION INJECTION

Shin Fujita

Shimadzu Asia Pacific Pte Ltd.

Keywords: phase space, periodic potential, Mathieu-Hillequation, eigen-trajectory

In the optimization of the quadrupole mass filter (QP filter), theunderstanding of ion motion in terms of the phase space (thecombined representation of the trajectory coordinate and mo-mentum) is useful. The phase space representation can give an‘ensemble’ behavior of ions inside the filter. Even though eachion trajectory does not have the RF periodicity of the appliedvoltages to electrodes, the phase space evolution does. It is onlywhen appropriate ensemble ions are considered together that aproper QP filter characterization is possible. We here report anew calculation framework for the phase space of the QP filter.The Mathieu-Hill equation is first solved for ‘complex number’eigen-trajectory that has pseudo RF periodicity (Bloch Function).It is then shown that the acceptance phase space can be derivedfrom Bloch Function without a need to calculate each ion trajec-tory. The ensemble behavior of ions can be estimated from oneBloch Function. The application of the Bloch Function methodto the pre-rod effects revealed that the ion injection efficiencymay significantly be improved by an appropriate choice of pre-rod (q, a) condition. Proper addition of DC voltage componentwill result in the phase space transformation in the pre-rod thatenables the efficient ion injection.

——————————————–

CS AND RB ION COLDBEAMS SUITABILITY FORCIRCUIT EDIT

Yuval Greenzweig, Roy M. Hallstein, Minh P. Ly, YarivDrezner, Rick H. Livengood, Shida Tan, and Amir Raveh

Intel Corporation

Keywords: Circuit Edit, Focused Ion Beam (FIB), Coldbeam,Cold FIB, Magneto-Optical Trap Ion Source (MOTIS), LowTemperature Ion Source (LoTIS)

Semi-conductor applications of Focused Ion Beams (FIB) havelong been enabled by Ga Liquid Metal Ion Source technology,but have been challenged by device density doubling every twoyears for the last 2-3 decades. One such application, Circuit Edit,has been critically losing performance due to density scaling in

3

the recent past. The emergence of several new FIB source tech-nologies such as gas field ionization and Coldbeam sources hav-ing much smaller probe sizes than Ga LMIS, and in some caseshigher secondary electron (SE) yields, promise a revival of FIBcapability for Circuit Edit. We report herein the results of ourtesting of Cs and Rb cold ion beams and their suitability for Cir-cuit Edit. Testing of the Cs coldbeam was performed at ZeroKNanotech, and Rb testing was performed at TU Eindhoven. Wecharacterized Cs image resolution, beam profile, minimum sizesof micro-trenches etched in SiO2, and material properties of Csdeposited dielectric and metalization. For both Cs and Rb wemeasured residual ion contamination levels and SE yields fromseveral common micro-electronic materials. We established thelack of invasiveness of Cs and Rb for Circuit Edit related oper-ations on 14nm Intel transistors. Lastly, using the Cs coldbeamwith gas chemistries for etching and deposition, we demonstratedthe first ever Cs based real Circuit Edits, which we performed onIntel 10nm chips.

——————————————–

ORBITRAP MASS SPECTROMETRY ANDNONLINEAR SPACE CHARGE DYNAMICS

D. Grinfeld, H. Stewart, M. Skoblin, E. Denisov, A.Makarov

Thermo Fisher Scientific, Bremen, Germany

Invited TalkKeywords: mass-spectrometry, space charge

The Orbitrap (TM) mass analyzer belongs to the family ofFourier transform mass spectrometers (FT MS) in which the ionsare trapped between two spindle-like electrodes. The quadro-logarithmic field [1,2] provides the quadratic effective potentialin the axial direction, in which the oscillatory frequency is in-dependent of the orbital parameters (e.g. the amplitude) so thations with same mass-to-charge ratios (m/z) preserve their com-mon phase in the course of 1e5 ÷ 1e6 oscillations. On the otherhand, ions with different m/z ratios oscillate with different fre-quencies, and Fourier analysis of the induced-current signal pro-duces a mass spectrum with a resolving power of up to one mil-lion [3]. If the number of injected charges is large, however, theion motion is affected by Coulomb forces. Though the space-charge field constitutes only ∼1e-4 of the trapping field, the in-teraction between ions with equal or close m/z is amplified underthe resonance conditions and results in sophisticated, sometimescounterintuitive, ion dynamics. For example, the Coulomb in-teraction between same charge ions, being repulsive by its na-ture, generates an effective attraction force. Detected FT peaksof ions with different but close m/z appear shifted towards eachother in frequency. Ultimately, it leads to complete frequencysynchronization, referred as coalescence, when the ions oscil-late as a single bunch. Coalescence makes the involved ionicspecies undistinguishable in the FT spectrum and thus deterio-rates the mass resolving capability. Basing on the perturbationtheory, we have developed a multi-body model and an algorithmcharacterizing ion dynamics in the presence of Coulomb interac-tions. Systems of multiple ions were simulated near and beyondthe coalescence threshold. It was shown that small non-idealities(∼1e-5) of the trap’s field substantially alter the ion dynamics,decreasing or increasing the threshold by a factor of ten. This

observation suggests approaches to improving space charge ca-pacity of the mass analyzers and, therefore, their dynamic range.

——————————————–

ELECTRON OPTICS OF A MULTI-BEAM SOURCE(MBS)

Ali Mohammadi-Gheidari, Xiaoli Guo and Pieter Kruit

Delft University of Technology, Lorentzweg 1, 2628CJDelft, The Netherlands

Keywords: Electron optics, Multi-beam inspection, Highthroughput, High resolution

In the field of Charged Particle Optics, more than 50 years ofresearch and development have been devoted to improving theresolution of these systems. Present day systems can easily ob-tain subnanometer resolution in imaging and sub 10nm resolu-tion in patterning. To preserve the high resolution the currentin the probe should be low, only tens of pico-Amps to a fewnano-Amps. This makes the throughput of these systems toolow for applications such as 3D imaging or wafer inspection.Multi-electron beam systems, in which not one but many elec-tron beams are focused onto the sample simultaneously, can en-hance the throughput to a great extent. A Multi-Beam ScanningElectron Microscope (MBSEM) that delivers 196 focused beamsonto the sample, based on an FEI Nova-Nano 200 SEM electronoptical column was designed and built at the CPO group in DelftUniversity of Technology[1]. The multi-beam source (MBS) forthis system, presented by Zhang et al.[2], was based on the prin-ciple of the “zero-strength” lens, where the deflection and col-limation of the off-axis beams take place in a conjugate planeof the source in order to avoid chromatic deflection aberrations,astigmatism and coma. In further theoretical investigations andsimulations, however, it was found that even if the net deflectionat the aperture array plane is zero, the axial spherical and chro-matic aberrations of the field in front of the aperture array planecan lead to unacceptable astigmatism and extra contributions tothe field curvature in the image plane of the array. Therefore, inthe design of the MBS, it is not the zero deflection at the aper-ture plane that should be respected but the minimization of itsspherical and chromatic aberration coefficients. In general, theaberrations of round electron lenses cannot be made arbitrarilysmall or negative. However, the unique property of the presentlens allows the aberrations to be made close or even equal tozero! A comprehensive and detailed analysis and design of theMBS will be presented here. [1] A. Mohammadi-Gheidari, C.W.Hagen and P. Kruit, JVST B 28(6) 2010. [2] Y. Zhang and P.Kruit, Physics Procedia 1 553, 2008.

——————————————–

4

DECELERATION OF HEAVY IONS, HITRAP ANDCRYRING@ESR

Frank Herfurth

GSI Helmholtz Centre for heavy ion research

Invited TalkKeywords: deceleration, highly-charged ions, low energy storagering, penning trap

To perform precision experiments on exotic ions, if highlycharged or rare, it is mandatory to provide means to link high-energy production schemes with low energy storage and mea-surement schemes. Only low energy storage in rings andtraps ensures the required extended observation time and wellcontrolled environment. At GSI/FAIR in Darmstadt/Germany,heavy, highly charged stable and rare ions up to bare uraniumare produced in large quantities. Medium charged ions at a few100 MeV/nucleon will be stripped of all electrons when sentthrough a thin foil or fragmented in nuclear reactions when in-teracting with enough material in thicker targets. The deceler-ation down to MeV/nucleon, keV/nucleon and finally sub meVrequires several steps involving storage rings and finally a dedi-cated linear decelerator coupled to ion traps. Two facilities, thelinear decelerator facility HITRAP and the low energy storagering CRYRING@ESR, will be introduced with planned experi-ments and status of installation.

——————————————–

MULTI-ELECTRODE LENS SYSTEMOPTIMIZATION USING GENETIC ALGORITHMS

N. Hesam Mahmoudi Nezhad1*, M. Ghaffarian Niasar2,A. Mohammadi-Gheidari1, T. Janssen3, C.W. Hagen1, P.

Kruit1

1 Delft University of Technology, Fac. Appl. Sciences,Dept. Imaging Physics, Charged Particle Optics Group,Lorentzweg 1, 2628 CJ Delft, The Netherlands; 2 Delft

University of Technology, Faculty of ElectricalEngineering, DC systems, Energy conversion and Storage,

Mekelweg 4, 2628 CD Delft, The Netherlands; 3 DelftUniversity of Technology, Faculty of Applied

Mathematics, Optimization, Mourik Broekmanweg 6,2628 XE, The Netherlands

Keywords: Electrostatic lens Optimization, Second OrderElectrode Method (SOEM), Genetic Algorithms (GAs)

In electron lens design, finding the optimum lens system for theapplication at hand, is still quite a challenge. The situation be-comes especially more complicated when many lens electrodesare involved, because the number of free parameters of the opti-mization, such as electrode thickness, radii, gaps between elec-trodes and voltages, increases rapidly. Therefore, fast optimiza-tion routines are needed to tackle the problem. In the past, therehave been some attempts to develop such optimization programs.Szilagy et al. [1] and Adriaanse et al. [2], have published someresults in 1989 on rough optimization of electrostatic lenses.However, using the above-mentioned methods, one could not get

very accurate results. Now that we have more powerful com-puters and significantly better software, we revisit the problem.First we applied the so called “SOEM” (Second Order ElectrodeMethod) [2] for a fast (∼0.1sec) calculation of the axial poten-tial. However, the results of the optimization were not accurateenough. To improve the accuracy of the SOEM-based optimiza-tion, we integrated a finite element based potential calculationmethod (using COMSOL). This way the potential calculationand the objective function calculation is more accurate, althoughthe optimization becomes much slower. We propose a new ap-proach that improves on the low speed of optimization whilekeeping the high accuracy results of the finite element methodbased potential calculation. This is done by first using a roughoptimization based on the SOEM approach, resulting in a fewapproximately optimized systems. Then, using the parametersof the systems found, new sets of systems were defined usinga small range of values around these parameters. Then the moreaccurate, COMSOL-based optimization was applied to this set oflimited systems. We have tested our method on multi electrodesystems up to 7 electrodes. We succeeded to very accurately op-timize these systems within a few hours, with the electrode radii,gaps and voltages as free parameters, and the focus position asa constraint. [1] M.Szilagi. Yakowitz and M. Duff, Appl. Phys.Lett. 44, pp. 7-9, 1984. [2] J.P. Adriaanse, H.W.G Van der Steenand J.E. Barth, J.Vac. Sci. Technol. B7, pp. 651-666, 1989.

——————————————–

CORRELATIVE MICROSCOPY BASED ONSECONDARY ION MASS SPECTROMETRY FOR

HIGH-RESOLUTION HIGH-SENSITIVITYNANO-ANALYTICS

Hung Quang Hoang, Jean-Nicolas Audinot, SanthanaEswara, Tom Wirtz

Advanced Instrumentation for Ion Nano-Analytics(AINA), MRT Department, Luxembourg Institute of

Science and Technology (LIST), 41 rue du Brill, 4422Belvaux, Luxembourg

Invited TalkKeywords: Correlative microscopy, secondary ion massspectrometry, helion ion microscopy, transmission electronmicroscopy, scanning probe microscopy

Nano-analytical techniques and instruments providing both ex-cellent spatial resolution and high-sensitivity chemical informa-tion are of extreme importance in materials science and life sci-ences for investigations at the nanoscale. New characterisationtools need to anticipate these research trends, but as more andmore techniques approach their fundamental limits it is only bycombining multiple techniques that disruptive advances may bemade. While techniques such as Electron Microscopy, HeliumIon Microscopy and Scanning Probe Microscopy are commonlyused for high-resolution imaging, they provide no or only lim-ited analytical information. In particular, both Energy-DispersiveX-Ray Spectroscopy (EDX) and Electron Energy Loss Spec-troscopy (EELS) that are routinely used in electron microscopyhave limited sensitivity, neither can distinguish isotopes and bothhave difficulty with light elements. In contrast, Secondary IonMass Spectrometry (SIMS) offers extremely high chemical sen-

5

sitivity, but it typically suffers from poorer lateral resolution.However, by combing SIMS with one of these high resolutionmicroscopy techniques, these intrinsic drawbacks may be over-come [1]. Therefore, in order to get chemical information witha highest sensitivity and highest lateral resolution, we developedintegrated instruments combining SIMS with Transmission Elec-tron Microscopy [2], Helium Ion Microscopy [3-5] and Scan-ning Probe Microscopy [6] and developed associated correlativemethodologies and workflows. These workflows allow TEM, SEand SPM images of exactly the same zone analysed with SIMSto be acquired easily and rapidly, followed by a fusion betweenthe SE and SIMS data sets [7]. In this talk, we will presentthe concepts, describe the instruments and discuss their perfor-mance characteristics. We will then present a number of ex-amples taken from various fields of materials science and lifescience to show the powerful correlative microscopy possibil-ities enabled by these new in-situ methods. [1] T. Wirtz, P.Philipp, J.-N. Audinot, D. Dowsett, S. Eswara, Nanotechnology26 (2015) 434001 [2] L. Yedra, S. Eswara, D. Dowsett, T. Wirtz,Sci. Rep. 6 (2016) 28705 [3] T. Wirtz, D. Dowsett, P. Philipp,Helium Ion Microscopy, edited by G. Hlawacek, A. Golzhauser,Springer, 2017 [4] D. Dowsett, T. Wirtz, Anal. Chem. 89 (2017)8957-8965 [5] P. Gratia et al, J. Am. Chem. Soc. 138 (49)15821-15824, 2016 [6] Y. Fleming et al., Beilstein J. Nanotech-nol. 6 (2015) 1091 [7] F. Vollnhals, J.-N. Audinot, T. Wirtz, M.Mercier-Bonin, I. Fourquaux, B. Schroeppel, U. Kraushaar, V.Lev-Ram, M. H. Ellisman, S. Eswara, Anal. Chem. 89 (2017)10702-10710

——————————————–

SPHERICAL ABERRATION CORRECTION WITHIN-LENS N-FOLD SYMMETRIC LINE CURRENTS

S. Hoque1,2), R. Nishi1), H. Ito1,3), A. Takaoka1)

1) Osaka University, 2) Hitachi High TechnologiesAmerica, Inc., 3) Hitachi High-Technologies Corp.

Keywords: Spherical aberration, Hexapole corrector, Sextupolecorrector, N-SYLC

We have shown that N number of line currents placed symmet-rically about the optic axis generate 2N-pole fields [1]. Wecall this structure N-fold symmetric line currents, or, N-SYLCin short. We have proposed simple aberration corrector mod-els based on N-SYLC for scanning electron microscopes (SEM)[2][3][4]. The most important feature of N-SYLC is that it isfree of magnetic material, thus in principle eliminates the prob-lems of hysteresis, non-uniformity, and magnetic saturation suf-fered by conventional magnetic multipoles. We have shown the-oretically that the conventional multipoles of sextupole doubletmodel of H. Rose [5][6] can be replaced with 3-SYLC to correctspherical aberration [2]. Here, we consider a new structure super-imposing N-SYLC on rotationally symmetric lens fields, whichis only possible because N-SYLC is free of magnetic materials.This simplifies the corrector structure, and allow for miniatur-ization and more versatile design. We call this structure in-lensN-SYLC. We show by analytical calculation that by adjustingcertain parameters of the system, in-lens 3-SYLC can generatenegative spherical aberration with high sensitivity, so that it canbe used to correct the spherical aberration of objective lens. Wealso verify the results by computer simulation [7]. [1] Nishi R.,

Ito H., Hoque S.: Wire corrector for aberration corrected elec-tron optics, (IMC2014), IT-1-P2984, pp.200-201. [2] S. Hoque,H. Ito, R. Nishi, A. Takaoka, E. Munro: Spherical aberration cor-rection with threefold symmetric line currents, Ultramicroscopy161, (2016) 74-82. [3] S. Hoque, H. Ito, A. Takaoka, R. Nishi,Axial geometrical aberration correction up to 5th order with N-SYLC, Ultramicroscopy 182, (2017) 68-80. [4] P. W. Hawkesand E. Kasper, Principle of Electron Optics, Volume 2: Ap-plied Geometrical Optics, second edition, chapter 41, p. 986-988, Academic Press, 2017. [5] H. Rose, Correction of apertureaberrations in magnetic systems with threefold symmetry, Nu-clear Instruments and Methods in Physics Research 187, 187-199(1981). [6] M. Haider, H. Rose, S. Uhlemann, B. Kabius, andK. Urban, Towards 0.1 nm resolution with the first sphericallycorrected transmission electron microscope, Journal of ElectronMicroscopy 47, 395 (1998). [7] Spherical aberration correctionwith an in-lens N-fold symmetric line currents model, Ultrami-croscopy 187, (2018) 135-143.

——————————————–

HIGH-PRECISION MASS MEASUREMENTSWITH MR-TOF-MS

Timo Dickel, Christine Hornung

Justus-Liebig-University, Giessen, Germany / GSIHelmholzcenter for Heavy Ion Research, Darmstadt,

Germany

Invited TalkKeywords: TOF-MS, RF traps, mass measurements

At the FRS Ion Catcher at GSI, projectile and fission fragmentsare produced at relativistic energies at the FRS, separated in-flight, range-focused, slowed-down and thermalized in a cryo-genic stopping cell and transmitted to a multiple-reflection time-of-flight mass spectrometer (MR-TOF-MS). The MR-TOF-MScan perform direct mass measurements of exotic nuclei, to pro-vide an isobarically and isomerically clean beam for further ex-periments, and as a versatile diagnostics device to monitor theproduction, separation and manipulation of exotic nuclei. At theJustus Liebig University, Giessen, Germany similar MR-TOF-MS have been developed for the ISOL facility TRIUMF, Van-couver, Canada and for applications in analytical mass spectrom-etry. These MR-TOF-MS consist of an injection RF trap to formthe ns ion bunches, a coaxial isochronous TOF analyzer, and aTOF detector for mass measurement and a Bradbury-Nielsen-Gate for spatial mass separation. Several novel principles fur-ther enhance the performance and versatility of the MR-TOF-MS, including (i) a post-analyzer reflector, (ii) a dynamical time-focus shift technique and (iii) mass-selective re-trapping in theRF trap. Thus extremely versatile MR-TOF-MS with mass re-solving powers beyond 600,000 (FWHM), high transmission ef-ficiency, ion capacities of more than a million ions per secondand cycle frequencies has high as 1kHz have been developed.The MR-TOF-MS can also be used as their own isobar separator.Mass measurements of uranium projectile and fission fragmentsproduced at the FRS at 1000 MeV/u have been performed usingthe MR-TOF-MS of the FRS Ion Catcher. More than 30 short-lived ground state masses have been measured with high massaccuracies (down to 6E-8). The excitation energies of isomersand isomeric ratios were determined using mass spectrometry,

6

and, for the first time, an isomeric beam was prepared using anMR-TOF-MS. The unique combination of performance parame-ters make the MR-TOF-MS the system of choice for measuringthe masses of very exotic nuclei and for the search for new long-lived isomeric states.

——————————————–

LATTICE DESIGN OF THE HEPS

Yi Jiao, Gang Xu, Qing Qin

Institute of High Energy Physics, CAS

Invited TalkKeywords: High Energy Photon Source, diffraction limitedstorage ring, linear optics, nonlinear dynamics

The High Energy Photon Source (HEPS) is the first high-energydiffraction-limited storage ring (DLSR) light source to be built inChina,with a natural emittance of a few tens of picometers and acircumference of 1360 m. After 10 years’ evolution, the accerla-tor physics design of the HEPS has been determined. The latestHEPS lattice consists of 48 hybrid-7BAs with a few modifica-tions, such as, antibends, superbends, and alternating high- andlow-beta sections. These modifications promises a 34 pm designwith high brightness. In this report we will introduce the statusof the HEPS acclerator physics design and the linear optics andnonlinear dynamics of the latest HEPS lattice.

——————————————–

DETECTION SYSTEMS IN SCANNINGELECTRON MICROSCOPE

Jaroslav Jiruse

TESCAN Brno

Keywords: SEM, detection system, secondary electrons,backscattered electrons

Controlling surface sensitivity is becoming increasingly impor-tant in SEM. We will present results obtained with ultra-highresolution columns developed recently with extended detectionsystems optimized for low energies. These systems allow angu-lar filtering of secondary electrons and both angular and energyfiltering of backscattered electrons. These filtering possibilitieslead to enhanced surface sensitivity of the detected signal.

——————————————–

DESIGN OF A HIGH-PERFORMANCEPOST-COLUMN IMAGING ENERGY FILTER FOR

(S)TEM INSTRUMENTS

Frank Kahl, Heiko Muller, Martin Linck, RichardSchillinger

CEOS GmbH, 69126 Heidelberg, Germany

Invited TalkKeywords: post-column energy filter, ESI, EELS,Electron-Spectroscopic-Imaging,Electron-Energy-Loss-Spectroscopy, spectroscopy

CEOS developed a high-performance post-column imaging en-ergy filter for transmission electron microscopy. It can be usedfor Electron-Spectroscopic-Imaging (ESI) and Electron-Energy-Loss-Spectroscopy (EELS) for beam energies from 30kV to300kV. It has been designed as a multi-purpose instrument forzero-loss filtering with large field of view, low-loss and core-loss spectroscopy with highest energy resolution, angle-resolvedEELS, and fast STEM-EELS mapping applications. Its excep-tionally low remanence and drift effects allow for switching fromhigh-resolution EELS to ESI or another EELS dispersion backand forth with only very little shift or defocus. The excellentstability of the filter supply minimizes the need for regular re-tuning. The filter is supported by a python-based image process-ing platform featuring automated tuning for maximum ease ofuse. The users can run their own python scripts, having accessto all acquired images plus image processing and display func-tionality via an Application-Programmers-Interface (API). Thirdparty software can be integrated very easily. The detector in-terfaces are designed for flexibility. In ESI mode an entranceaperture of up to 12 mm can be used. The filter supports post-filter cameras with detector sizes up to 64 mm side length andfits into the mounting space of all modern (S)TEMs. Three ESImagnifications are supported, imaging a quadratic field of viewsof 8 mm, 10 mm and 12 mm diagonal length in the entranceaperture plane onto the detector. For those three magnificationsthe non-isochromaticities, maximum distortions and maximumchromatic distortions are (0.13 eV, 0.9 %, 0.22 %), (0.44 eV, 1.0%, 0.23 %) and (1.7 eV, 1.4 %, 0.26 %), respectively, for an en-ergy window of 50 eV at 200 kV. The regular EELS aperture is 5mm. The pre-slit alignment is kept almost identical for the differ-ent ESI and EELS modes. Additional spectroscopy modes sup-porting an extremly large spectral range of up to 4 keV at 200 kVand dedicated alignments for omega-q mapping at higher disper-sions are possible. For highest energy resolution a dispersion of 2meV/channel for 4k x 4k detectors with pixel sizes at the order of15 um is available. In order to minimize the non-isochromaticityof an ESI image and to improve the quality of focus of the Zero-Loss-Peak (ZLP) in EELS mode, the diameter of the ZLP in theslit plane formed by all electrons over the entire entrance aper-ture must be as small as possible in the dispersive direction. Thisis equivalent to minimize the geometric aberrations in the slitplane, which is achieved by combining a sector magnet whosegeometry has been optimized for minimum intrinsic geometricaberrations with a set of sextupole, octupole, decapole and dode-capole fields allowing for correcting all residual geometric slitaberrations up to third order and certain aberrations of fourthand fifth order. The aberrations are measured by scanning theZLP over the edge of one selection slit, recording the attenua-tion pattern and fitting the aberration coefficients from that. The

7

pre-slit setting shared by all ESI magnifications is optimized forlow residual chromatic distortions in the image while the pre-slitsetting shared by all regular EELS modes uses an additional pre-slit sextupole to correct for the spectrum inclination. For all ESImagnifications the projective corrects or adjusts the distortion co-efficients of first order, the three intrinsic coefficients of secondorder, one critical intrinsic coefficient of third order and finallythe chromatic distortion coefficients of first degree. Its excel-lent performance is achieved by an optimized design comprisinga minimalistic design of only four main quadrupoles, two weakrotated quadrupoles for correcting parasitic aberrations and threesextupoles. The measureable residual distortions for the 12 mmentrance field of view are very predictable and can optionally beremoved by online distortion dewarping.

——————————————–

ON SIGNIFICANCE OF 50 KV E-BEAM SHOTNOISE IN LITHOGRAPHY APPLICATION

Victor Katsap

NuFlare Technology America

Keywords: e-beam lithography, shot noise, acid deficit

In e-beam mask writers, 50-keV e-beam having dynamicallycontrolled rectangular shape strikes mask blank surface. Maskblank consists of thin layers of resist and metal atop of massivequartz substrate. Primary and secondary electrons expose re-sist by colliding inelastically with resist molecules. Beam dwelltime depends on beam spot current density and resist sensitivity.Noise is often defined as ratio of SQRT(var) to MEAN value.For e-beam shot noise, it’s 1/SQRT(Ne), Ne being number ofelectrons in process considered. In a simplified way, e-beamresist exposure may be looked at as a sequence of 3 steps: Neelectrons enter resist > Na acids get de-protected > Np polymermolecules get exposed in Post-Exposure-Baking (PEB) step. Ineach step, limited number of particles is involved, and so eachprocess may be described with Poisson-type statistics. These 3statistics can be summed up to evaluate resulting noise in e-beamexposed and developed resist: SQRT(var)/MEAN = SQRT(1/Ne+ 1/Na + 1/Np) This means that step with the least number of par-ticles would dominate resulting noise in the exposed feature, thusdefining exposure quality. Typically, Na is smallest of 3 values,facilitating so called “acid deficit”. Under realistic conditions, e-beam noise is 1/4 of total, while acid noise may reach 70% of to-tal noise. This could explain a difference in exposing resist withlow- and high-energy e-beams: - with low-energy beam, acidyield is close or greater than 1 acid per primary electron, henceno acid deficit, and e-beam noise dominates. - with high-energybeam, acid yield is less than 1 acid per primary electron, henceacid deficit, and e-beam noise is minor factor. However, becauseof using thinner and thinner resist in commercial process, e-beamnoise may become major component of the total exposure noise.

——————————————–

HIGH-ORDER ABERRATIONS OF LARGEAPERTURE MAGNETS AND APPLICATIONS TO

THE SUPER-FRS PROJECT AT GSI

Erika Kazantseva, Oliver Boine-Frankenheim, HelmutWeick

Technische Universitaet Darmstadt

Invited TalkKeywords: Super-FRS, high-order aberrations, realistic Taylortransfer maps

The magnets of the charged particle spectrometers and separa-tors play a decisive role in the beam quality and transmissionpercentage, especially for the systems with large geometricaland momentum acceptances. In the case of the Superconduct-ing Fragment Separator (Super-FRS), a core part of the FAIRproject being built at GSI, the undesired high-orders aberrationsare expected due to the large usable apertures of the magnets(38x19 cm2 in the main dipoles and 19x19 cm2 in the multipoles)and wide operation rigidity range of 2-20 Tm. In this work wewill analyse the aberrations introduced by the normal conduct-ing dipole magnet of the Super-FRS preseparator. The methodsof generating the high-order Taylor transfer maps from the 3Dmagnetic field distributions and taking the magnetic rigidity intoaccount will be discussed.

——————————————–

QUANTUM CHARGED-PARTICLE BEAM OPTICS

Ramaswamy Jagannathan and Sameen Ahmed Khan

Chennai Mathematical Institute and Dhofar University

Keywords: Quantum Mechanics, Charged Particle Beam Optics

Though the classical charged-particle beam optics is very suc-cessful, in designing and operating numerous charged-particlebeam devices from electron microscopes to particle accelerators,it is natural to look for the quantum theory of such systems han-dling beams of microscopic particles for which quantum me-chanics should be relevant. With the curiosity of understand-ing how the classical charged-particle beam optics is so success-ful, the quantum charged-particle beam optics (QCPBO) is be-ing developed by Jagannathan et al. QCPBO is seen to repro-duce the classical charged-particle beam optics exactly for theparaxial and aberrating systems in the classical limit of drop-ping the additional quantum correction terms which depend onthe Planck constant and are, of course, extremely small com-pared to the classical terms. In the classical limit the quan-tum formalism reproduces the well-known Lie algebraic formal-ism of classical charged-particle beam optics. QCPBO basedon the Klein-Gordon equation is applicable to spin-0 and spin-less particles. The formalism of QCPBO based on the Diracequation provides a unified treatment of orbital and spin dy-namics of a Dirac particle with anomalous magnetic momentbeing transported through magnetic optical elements accountingfor the orbital phase-space transfer maps, including the Stern-Gerlach effect, and the Thomas-Bargmann-Michel-Telegdi spinmotion. QCPBO based on the nonrelativistic Schrodinger equa-tion emerges as the approximation of the relativistic formalismsbased on both the Klein-Gordon and the Dirac equations.

8

——————————————–

A LOW-ENERGY SPREAD GRAPHENE COATEDNICKEL ELECTRON SOURCE FOR

LOW-VOLTAGE SCANNING ELECTRONMICROSCOPY

Xiuyuan Shao, Wei Kean Ang, and Anjam Khursheed

Department of Electrical and Computer Engineering,National University of Singapore, 4 Engineering Drive 3,

Singapore 117583, Singapore

Keywords: Energy distribution, Boersch effect, Graphene

Field emission energy distribution characteristics of graphenecoated nickel emitters have been experimentally measured as afunction of cathode extraction voltage and tip radius. These emit-ters have been recently reported to have significant advantagesover conventional cold field tungsten emitters for electron mi-croscopy/lithography applications. Full-width at half-maximumvalues for their energy spectra were experimentally measured tovary between 0.16 to 0.39 eV. By subtracting the calculated to-tal energy distribution (TED) in the thermal field regime fromthe experimental spectra, the energy spread broadening due toCoulomb interactions (Boersch effect) was obtained, and foundto increase with increasing extraction voltages and decreasingthe tip sizes. These results are of particular interest for low-voltage scanning electron microscopy applications where chro-matic aberration of the objective lens is the main factor limitingspatial resolution.

——————————————–

CONSIDERATIONS ON BEAM OPTICS OFSUPERCONDUCTING HEAVY ION LINACS FOR A

RARE ISOTOPE BEAM FACILITY

Jongwon Kim, Brahim Mustapha

Institute for Basic Science, Korea, Argonne NationalLaboratory, USA

Invited TalkKeywords: Superconducting linac, Heavy-ion beam, Rareisotope beam

The rare isotope science project (RISP) was started in Korea fromDec. 2011 to establish an accelerator facility based on supercon-ducting linacs for nuclear and applied science studies under theauspices of the Institute for Basic Science (IBS). The current de-sign was frozen in 2013 in terms of the facility layout and thecivil construction began in 2016. In fact, some considerations onalternative linac design were made in 2016 together with the linacdevelopment group of Argonne National Lab in search of furtheroptimized configuration of the driver linac, which should havesound lattice design against realistic machine errors. Results ofbeam optics simulations and error analysis for an alternative de-sign will be presented. Also, considerations on optimized designand operation scheme of the superconducting linac which can ac-celerate both stable and rare isotope beams simultaneously, willbe presented.

——————————————–

PONDEROMOTIVE GENERATION ANDDETECTION OF ATTOSECOND ELECTRON

PULSES

Martin Kozak, Timo Eckstein, Norbert Schonenberger,Peter Hommelhoff

Faculty of Mathematics and Physics, Charles University,Ke Karlovu 3, 12116 Prague 2, Czech Republic,

Department of Physics, Friedrich-Alexander-UniversitatErlangen-Nurnberg (FAU), Staudtstrasse 1, 91058

Erlangen, Germany

Invited TalkKeywords: Ultrafast control of electron pulses, attosecondphysics

In this contribution we will discuss recently developed techniquefor generation and detection of attosecond electron pulses via in-elastic ponderomotive scattering of electrons at an optical trav-eling wave formed by two laser pulses at different frequencies.This scheme represents an analogy to the classical Kapitza-Diraceffect [1], in which the roles of the transverse and longitudinal di-rections (with respect to electron propagation) are reversed. Wedemonstrate a large modulation of the kinetic energy of subrela-tivistic electrons with initial kinetic energy of 29 keV, achieving apeak acceleration gradient of G=2.2 GV/m (energy gain/travelleddistance) [2]. A time-correlated modulation of electron energyleads to a ballistic compression and formation of attosecond elec-tron pulses. Detection of the sub-cycle temporal structure of theelectron pulse train was performed via energy streaking using asecond phase-controlled traveling wave [3]. Measured spectro-grams (spectrum as a function of relative phase) and their com-parison with numerical calculations allow monitoring the evolu-tion of the electrons’ longitudinal phase space distribution. [1] P.L. Kapitza, and P. A. M. Dirac, Proc. Cambridge Philos. Soc. 29,297 (1933). [2] M. Kozak, T. Eckstein, N. Schonenberger, andP. Hommelhoff, Nat. Phys. 14, 121-125 (2018). [3] M. Kozak,N. Schonenberger, and P. Hommelhoff, Phys. Rev. Lett. 120,103203 (2018).

——————————————–

CORRECTION OF ABERRATIONS IN ELECTRONMONOCHROMATORS AND SPECTROMETERS

N. Dellby1, A.L. Bleloch1, M.V. Hoffman1, T.C.Lovejoy1, C. Su2 and O.L. Krivanek1,3

1. Nion R&D, 11511 NE 118th St, Kirkland, WA 98034,USA 2. Nuclear Science and Engineering Department,MIT, Cambridge, MA 02139, USA 3. Department ofPhysics, Arizona State University, Tempe, AZ 85287,

USA

Invited TalkKeywords: spectrometers, monochromators, aberrationcorrection

Electron energy loss spectroscopy (EELS) in the electron mi-croscope has progressed remarkably in recent years. In ourmonochromated microscope system, we have reached 4.2 meV

9

energy resolution at 30 keV, measured as the full-width at half-maximum of the monochromated zero loss peak (FWHM ofZLP), and 5.9 meV at 60 keV - i.e., energy resolution better than1 part in 107. In both the monochromator and the spectrome-ter, the energy spectrum contains a small image of the electronsource, and the attainable energy resolution is given by ER = d /D, where d is the diameter of the source image and D the energydispersion. Reaching better energy resolution thus requires mak-ing d smaller, or increasing D. The second path employs sizableenergy-dispersing devices with concomitant weight and mechan-ical stability issues. We have chosen the more practical first pathto better energy resolution, and we pay particular attention tothree performance aspects: 1) stability of the source image po-sition, 2) demagnifying the source sufficiently (while retaining auseful electron current), and 3) increasing the convergence an-gle, to minimize the diffraction limit, and to maximize the probecurrent. The third aspect requires that aberrations be kept undertight control. Our latest spectrometer design corrects all geomet-ric aberrations up to 5th order, plus mixed chromatic-geometricaberrations up to third rank. It also includes autotuning that mea-sures individual aberration coefficients and corrects them. Thetalk will review the methods we use to optimize performance as-pects 1) and 2), and then focus on the correction of aberrationsin the monochromator and the spectrometer. It will also describeour recent work on optimizing probe corrector performance.

——————————————–

DOUBLE MIRROR ABERRATION CORRECTOR

Pieter Kruit(1) and Hideto Dohi(2)

(1)Delft University of Technology and (2)HitachiHigh-Technologies Corporation

Keywords: scanning electron microscope; aberration corrector;MEMS electron optical components

The resolution of scanning electron microscopes (SEMs) is lim-ited by aberrations of the objective lens, mainly the chromaticaberration. It is well known that both spherical and chromaticaberrations can be compensated by placing an electron mirrorin the beam path before the objective lens. The effectiveness ofthis has been proven in LEEM systems. Nevertheless, this ap-proach has not led to use of these aberration correctors in SEMs,probably because aberrations of the bending magnet can be a se-rious problem. We have proposed a mirror corrector with twomirrors placed perpendicularly to the optic axis of an SEM andfacing each. As a result, only small-angle magnetic deflectionis necessary to guide the electron beam around the top mirrorto the bottom mirror and around the bottom mirror to the objec-tive lens. The deflection angle is only in the order of 50 mrad,and thus sufficiently small to avoid deflection aberrations. In ad-dition, lateral dispersion at the sample plane can be avoided bythe correct choice of deflection fields. In order to keep such acorrector system simple, the incoming beam should pass the topmirror at a distance in the order of millimeters. It is proposedthat condition can be satisfied with micro-scale electron opticalelements fabricated using MEMS technology. Extensive opticalcalculations were performed. Aberrations of the micro-mirrorswere analyzed by numerical calculation. Dispersion and aberra-tions of the deflectors were calculated by using an analytical fieldmodel. We concluded that the proposed corrector system could

be a candidate for aberration correction in low-voltage SEMs.We have started the construction of a system to be tested in anexisting SEM. Reference: Hideto Dohi and Pieter Kruit, Designfor an aberration corrected scanning electron microscope usingminiature electron mirrors, Ultramicroscopy 189 (2018) 1-23.

——————————————–

SOME DESIGNS FOR QUANTUM ELECTRONMICROSCOPY

Pieter Kruit and Maurice Krielaart

Delft University of Technology, Lorentzweg 1, 2628CJDelft, The Netherlands

Keywords: interaction-free measurement, quantum electronmicroscope, electron mirror, electron phase manipulation

Following a recent suggestion [1] that interaction-free measure-ments may be possible with electrons, we have analyzed the op-portunities to use this concept for imaging of biological specimenwith reduced damage in a Transmission Electron Microscope.This requires that part of the electron wave travels multiple timesthrough the same position on the specimen. We expect this tobe an interesting challenge in charge particle optics. We havemade preliminary designs for an atomic resolution interaction-free electron microscope, or “quantum electron microscope” [2].The designs require a number of unique components not foundin conventional transmission electron microscopes. These com-ponents include a coherent electron beam-splitter or two-state-coupler, and a resonator structure to allow each electron to inter-rogate the specimen multiple times. A two-state-coupler has thefunction of moving the electron wave slowly between the ref-erence beam and the specimen beam, as in a Rabi-oscillation.We have suggested and are now investigating, both in simulationand in experiment, whether an electron mirror with a diffractiongrating on the reflecting surface can accomplish the two statecoupling while at the same time forming part of the resonator.The experimental design consists of many MEMS elements. Thesame mirror unit might be used for a more advanced manipula-tion of the electron wave front. This research is funded by theGordon and Betty Moore Foundation. 1] Putnam, W.; Yanik, M.Phys. Rev. A 2009, 80, 040902. 2] Kruit, P.; R. G. Hobbs, C-S. Kim, Y. Yang, V. R. Manfrinato, J. Hammer, S. Thomas, P.Weber, B. Klopfer, C. Kohstall, T. Juffmann, M. A. Kasevich,P.Hommelhoff, K. K. Berggren. Ultramicroscopy (2016), 31-453] Elitzur, A. C.; Vaidman, L. Found. Phys. 1993, 23, 987-997.

——————————————–

10

NANO APERTURE ION SOURCE

Leon van Kouwen, Dustin Laur and Pieter Kruit

Delft University of Technology, Lorentzweg 1, Delft, TheNetherlands

Keywords: Ion source; brightness; mirror charge

Ion production in the Nano Aperture Ion Source (NAIS) is basedon electron impact gas ionization inside a sub-micron sized gaschamber [1]. An important part of recent efforts [2] was devotedto understanding how the relevant physical processes determinethe ion beam performance. This has led to interesting insightsin charged particle optics. The influence of initial velocity andposition distributions of the neutral gas particles, their ioniza-tion cross sections, the electron current density, ion-neutral scat-tering, Coulomb interactions and the electric fields around thedouble membrane structure are studied by analytical models, nu-merical calculation, and ray tracing. An unexpected effect is thatthe low energy ions are deflected by their mirror charge whenthey exit the submicron sized orifice of the gas chamber. Animportant finding is that the ion current and the brightness tendto keep increasing with increasing particle density, despite in-creasing ion-neutral scattering. Ion-to-ion Coulomb repulsion isfound to pose a final limit to the achievable brightness. In a real-istic configuration, the simulations predict a brightness of about3 x 106 A/m2srV in combination with an energy spread of 1eV. In experiments we have now demonstrated a brightness of 1x 105 A/m2srV, which we consider a milestone result becauseit is already a competitive brightness when compared to a GaLMIS while there is clearly room for improvement. [1] David S.Jun, Development of the Nano-Aperture Ion Source, PhD ThesisTU Delft, 2014. [2] Leon van Kouwen, The Nano-Aperture IonSource, PhD Thesis TU Delft, 2017

——————————————–

LONGITUDINAL PHASE SPACE MANIPULATIONOF ELECTRON BEAMS USING MICROWAVE

CAVITIES

Wouter Verhoeven, Jasper van Rens, Peter Mutsaers, JomLuiten

Eindhoven University of Technology

Invited TalkKeywords: dynamic electron optics, ultrafast electronmicroscopy, microwave cavities, time-of-flight eels

At Eindhoven University of Technology we are developing res-onant microwave cavities as dynamic charged particle optics forelectron microscopy (EM). We employ miniaturized and power-efficient dielectric 3 GHz pillbox deflection cavities in TM-110mode both for creating femtosecond electron pulses by choppinga continuous beam and for measuring pulse lengths by streakingelectron pulses across a detection screen. Cavities in TM-010mode are used as longitudinal lenses to both compress electronpulses (positive focal length), to improve the temporal resolu-tion, or stretch them (negative focal length), to reduce the uncor-related energy spread. If properly used, microwave cavities donot affect the electron beam quality. The microwave phase can

be accurately synchronized to femtosecond lasers, enabling ul-trafast pump-probe experiments. Combining a high-quality con-tinuous electron gun with a special configuration of two TM-110cavities and two TM-010 cavities, time-of-flight EELS can be re-alized with few-10-meV energy resolution and few-ps time res-olution. This new method does not require femtosecond lasersand may constitute an interesting alternative to magnetic elec-tron spectrometry. The use of microwave cavities is therefore notrestricted solely to the burgeoning field of ultrafast EM. In fact,it may become increasingly relevant to EM in general.

——————————————–

FOR THE PROMOTION OF MICROSCOPY ANDMICROANALYSIS IN ALL RELEVANT

SCIENTIFIC DISCIPLINES: THE MICROSCOPYSOCIETY OF AMERICA

Charles E. Lyman and Robert L. Price

Dept. of Materials Science and Engineering, LehighUniversity, Bethlehem, PA (1991 MSA President) and

University of South Carolina Medical School, Columbia,SC (2018 MSA President)

Keywords: microscopy society, scientific conference, electronand ion optics, peer-reviewed journal, education

The Microscopy Society of America (MSA) is pleased to supportstudent travel to the 10th International Conference on ChargedParticle Optics. Such support is a natural fit for MSA given ourmission statement, an excerpt of which forms the title above.Founded 76 years ago, our Society has been a forum for dis-cussion of electron optics, from the early days of the TEM toaberration-correction technology; indeed, the original name ofthe society was the Electron Microscope Society of America.Today our purview also includes ion optics with presentationsand publications that include focused ion beam (FIB) and atomprobe tomography (APT) instrumentation and techniques. TheSociety promotes microscopy and microanalysis in several ways.The Microscopy & Microanalysis (M&M) conference in Augustis the largest annual microscopy meeting and instrument exhi-bition in the world. M&M 2018 featured 1236 scientific papersand attracted attendees from 40 countries. Among the many stu-dents and post-doctoral scholars attending the meeting, 55 re-ceived competitive travel awards. Our student council membersare learning leadership skills by independently organizing an an-nual pre-meeting congress. Microscopy and Microanalysis, ourpeer-reviewed research journal covering both the life and phys-ical sciences, ranks high among microscopy journals in termsof Impact Factor and number of pages published per year. Ourtechnical trade magazine, Microscopy Today, reaches a large au-dience and provides news and research summaries that empha-size advances in instrumentation and methods, as well as hintsand tips for novice and experienced microscopists. In additionto the above, our educational efforts include short courses andtutorials at the M&M meeting, a technologist certification pro-gram, undergraduate research scholarships, and several outreachprograms for students in middle school and high school. Finally,membership in MSA is cost-effective since dues cover subscrip-tions to our publications and registration discounts at the M&Mmeeting.

11

——————————————–

ELECTRON OPTICS FOR A MULTI-PASSTRANSMISSION ELECTRON MICROSCOPE

Marian Mankos1, Stewart A. Koppell2, Brannon B.Klopfer2, Thomas Juffmann3, Vladimir Kolarik4,

Khashayar Shadman1 & Mark A. Kasevich2

1 Electron Optica, 1000 Elwell Court #110, Palo Alto,California 94303, USA 2 Physics Department, StanfordUniversity, 382 Via Pueblo Mall, Stanford, California

94305, USA 3 Universitaet Wien, Campus ViennaBiocenter 5, 1030 Vienna, Austria 4 Delong Instruments,

Palackeho trida 3019/153b, Brno, Czech Republic

Invited TalkKeywords: electron optics, multi-pass transmission electronmicroscopy, electron mirror

Recent advances in cryo-electron microscopy (cryo-EM) and di-rect electron detection have spurred renewed interest in the de-velopment of novel electron imaging techniques for applicationsin structural biology at atomic resolution. The challenge withimaging unstained biological specimens is that they provide alow scattering cross-section to the probing electrons because theyare composed primarily of low atomic number elements. Hence,high electron doses are needed to obtain sufficient signal-to-noiseratios (SNR). Such doses, however, severely damage the spec-imens. Multi-pass transmission electron microscopy [1] is apromising approach that can reduce the required electron dosefor a desired SNR by exploiting the change to the phase of theelectron wave that is imparted by the specimen. In this approach,the electron beam interacts elastically with the specimen multipletimes so that the change in the phase accumulates before reach-ing the detector. Here we examine the electron-optical design ofa practical implementation of a multi-pass transmission electronmicroscope (MTEM), which is currently under construction. InMTEM, an electron pulse, triggered by an ultrafast laser beam,is focused by the illumination optics and transmitted by the en-trance electron mirror, rendered transparent by a voltage pulsesynchronized with the laser beam. The transmitted electron pulseenters an electron resonator, bounded by the entrance and exitmirrors. The resonator includes two objective and field lensesthat sandwich the specimen. The electron pulse is collimated bythe upper field and objective lens onto the specimen, and refo-cused by the lower objective and field lens onto the exit mirror,which reflects it back symmetrically so that the electron pulseis collimated again at the specimen. This reflection is carriedout multiple times until a second voltage pulse renders the exitmirror transparent to allow the electron pulse with the accumu-lated phase to proceed into the projection optics, which magnifiesthe image at the exit mirror onto the detector. Past simulationshave predicted an improvement in resolution and sensitivity for arange of electron microscopy imaging techniques, and an order-of-magnitude reduction in damage at equivalent resolution[2]. 1T. Juffmann, B. B. Klopfer, T. L. Frankort, P. Haslinger, and M.A. Kasevich, Nat. Commun. 7, 12858 (2016). 2 T. Juffmann, S.A. Koppell, B. B. Klopfer, C. Ophus, R. M. Glaeser, and M. A.Kasevich, Sci. Rep. 7, 1699 (2017).

——————————————–

ULTRAFAST ELECTRON DIFFRACTION USINGTHE CBETA PHOTOINJECTOR

Jared Maxson, William Li

Cornell University

Invited TalkKeywords: UED, photoinjector, time resolved diffraction

In this talk, I will describe perspectives and simulations of thephotoelectron injector for the CBETA accelerator (Cornell-BNLEnergy Recovery Test Accelerator) as a high average brightnesssource of short-pulse electrons for femtosecond time-resolvedelectron diffraction. Specifically, I will discuss the ability ofthe CBETA photoelectron gun to take advantage of novel highcoherence photocathode materials, as well as the extreme flex-ibility of the RF acceleration system. The acceleration systemis composed of five superconducting cavities each with indepen-dent phase and amplitude control, which I will show is critical fora number of applications, including longitudinal phase space lin-earization in velocity bunching, as well as ultrafast time of arrivalmeasurements when used in conjunction with a high resolutionspectrometer. These capabilities make the CBETA photoinjectoran excellent candidate for use in high spatiotemporal resolutionelectron scattering experiments.

——————————————–

COLD ATOM ION SOURCES

J. J. McClelland(1), J. R. Gardner(1,2), W. R.McGehee(1), A. Schwarzkopf(3), B. Knuffman(3), and A.

V. Steele(3)

(1) CNST, National Institute of Standards and Technology,Gaithersburg, MD, USA; (2) IREAP, University of

Maryland College Park, College Park, MD, USA; (3)zeroK NanoTech, Gaithersburg, MD, USA

Invited TalkKeywords: ion sources; laser cooling; high brightness; focusedion beams

Ionization of laser-cooled atoms has emerged as a new approachto constructing high brightness ion sources for applications suchas focused ion beam (FIB) microscopy and milling. While con-ventional sources, such as the Ga liquid metal ion source (LMIS)or the gas field ionization source (GFIS), attain brightness byemitting from a very sharp tip, cold atom sources reach highbrightness through reducing the transverse velocity spread. Withthe ultracold, microkelvin-range temperatures achievable withlaser cooling, the corresponding velocity spread can lead to abrightness significantly higher than typical LMIS values. More-over, the phase-space shape of the emittance of the source - nar-row in velocity, wide in space - brings new opportunities for ionoptical design. For example, high currents can be obtained with-out the high current density present in sharp tip sources. Thiscan result in fewer Coulomb effects, such as increased emittanceand broadened energy spread (Boersch effect). In addition, theabsence of a sharp tip eliminates a sensitivity to source stability.Other advantages of this type of source include insensitivity tocontamination, access to new ionic species, inherent isotopic pu-rity, and fine control over emission, down to the single ion level.

12

To date, sources have been demonstrated with Cr, Li, Rb, and Csions. In this talk I will review progress in the field, focusing onour work with a Li FIB microscope for battery studies and a CsFIB with brightness 24 times higher than the LMIS.

——————————————–

A NEW SIMULATION PROGRAM FORELECTRON MIRRORS USING THE BOUNDARY

ELEMENT METHOD

Eric Munro, John Rouse, Haoning Liu and CatherineRouse

Munro’s Electron Beam Software Ltd, 14 CornwallGardens, London SW7 4AN, United Kingdom

Keywords: Electron mirrors, boundary element method,aberration computation, direct ray tracing

Electron mirrors have many applications in charged particle op-tics, e.g. aberration correctors, beam separators, time-of-flightspectrometers. Optimization of electron mirrors requires highaccuracy computation of electric fields, trajectories, focal prop-erties, and geometrical and chromatic aberrations, including bothspatial and temporal aberrations. The simulations are more dif-ficult than for electron lenses, because the ray slopes becomeinfinite at the reflection plane, so the trajectories and aberrationshave to be computed using time, rather than axial position, asindependent variable. This paper present a new electron mirrorsimulation software using the Boundary Element Method. Thismethod computes the electric charge distribution generated onthe electrode surfaces, by dividing the electrodes into rotationallysymmetric rings, and generating a matrix equation expressing thepotential on each ring as a weighted sum of the surface chargeson all the rings. Since the potentials on each ring are the knownelectrode potentials, a matrix equation is obtained whose solutionyields the ring charges. The potential at any point in the mirrorcan then be computed by numerical evaluation of a sum of ellip-tic integrals of the ring charges. This method has several beauti-ful features: (1) Data input is simple, because only the electrodesurfaces need to be discretized, not the intervening space; (2)Numerical accuracy can be verified by evaluating the potentialat check-points on the electrodes and comparing these with theknown values; (3) The axial potential distribution and its deriva-tives can be obtained with great accuracy; (4) From these axialfunctions, all the optical properties can be computed, includinggeometrical and chromatic aberrations, both temporal and spa-tial, using a differential algebraic method; (5) The potential andfields at any off-axis point can be computed using elliptic inte-grals which have great accuracy and stability, and the results arean exact solution of Laplace’s equation; (6) These off-axis fieldscan be used to compute aberrations by direct ray-tracing, provid-ing an independent check on the differential algebra results. Theprogram will be described in detail, and illustrated with practicalexamples, and the results compared with those from our previouselectron mirror program, which used finite element method andHermite series fits.

——————————————–

ELECTRON RAY TRACING IN A CYLINDRICALDEFLECTOR ANALYZER FOR FIELD EMISSION

SPECTROSCOPY

Hidekazu Murata, Seiji Watanabe, Takahiro Ikeda,Hirotaka Asai, Eiji Rokuta, and Hiroshi Shimoyama

Faculty of Science and Technology, Meijo University

Keywords: electron ray tracing; three-dimensional boundarycharge method; Runge-Kutta-Fehlberg 4th order method; fieldemission spectroscopy; cylindrical deflection angle

Field-emission spectroscopy (FES) is a technique that acquiresan energy spectrum of the electrons emitted from a field emitter.A cylindrical deflector analyzer (CDA) is often used for FES.Assuming that the electric field in the CDA is an ideal cylindri-cal field, it is known that the optimum deflection angle is 127◦.In fact, there is a deviation from an ideal cylindrical field in theCDA. In particular, a fringing field occurring at the vicinity ofentrance and exit of slits is not negligible. Therefore, herein, weperformed three-dimensional (3D) electric field calculations andelectron ray tracing using a 3D boundary charge method (BCM)that we developed previously. Furthermore, for this purpose, weimproved the calculation method of the electron ray tracing. Sofar, we had used Runge-Kutta-Gill method (RKG method) for theelectron ray tracing. In the method, the step size is fixed due tothe specification of the method. Therefore, to perform an electronray tracing with high-accuracy, the step size must be finer. How-ever, extremely-long calculation time is required for the elec-tron ray tracing. Hence, to improve this obstacle, we introducedRunge-Kutta-Fehlberg 4th order method (RKF4 method) into theelectron ray tracing calculation so that suitable step size can becontrolled automatically while keeping high-accuracy. As a re-sult, we have found that the improved method can reduce thecalculation time while keeping high-accuracy. In addition, as aresult of the election ray tracing in the CDA, we also have foundthat when the potential difference between the inner and outerelectrodes is 0.353 V (= 2 E0/e log 1.8) and the initial energy E0of the electrons is 0.3 eV, the optimum deflection angle of theCDA is ∼109 degree.

——————————————–

LARGE MOMENTUM ACCEPTANCE BEAMOPTICS OF A SUPERCONDUCTING GANTRY

FOR PROTON THERAPY

K. P. Nesteruk, C. Calzolaio, A. Gerbershagen, D. Meer,V. Rizzoglio, M. Seidel, and J. M. Schippers

Paul Scherrer Institut, Villigen PSI, Switzerland

Keywords: Beam optics design; proton therapy; superconductinggantry

In proton therapy, the last part of the beam transport system isinstalled on a rotatable gantry, so that the beam can be aimedat the tumor from different angles. Since such a gantry systemconsists of many dipole and quadrupole magnets, it is typicallya 200 ton device of more than 10 m in diameter. The use of su-perconducting (SC) magnets for proton therapy allows gantries

13

to be significantly lighter and potentially smaller, which is at-tractive for this medical application. In addition to that, SC com-bined function magnets enable beam optics with a very large mo-mentum acceptance. The latter can be advantageous for patienttreatment, since the irradiation time can be significantly reducedby avoiding magnet current changes. A new prototype of a SCgantry with a momentum acceptance of +/- 15 % is under de-velopment at PSI. To design such an achromatic system, precisehigh-order calculations have been performed. In order to reachthe required accuracy and to check consistency of the obtainedresults, we have used several simulation tools in our iterative de-sign approach. Here we will describe how we have combinedan initial standard first order calculation with more detailed cal-culations using the higher order code COSY Infinity and particletracking using OPAL (open source software from PSI) in 3D fieldmaps obtained from detailed magnet calculations performed inOPERA. A comparison of the results from the beam-optics cal-culations helped to determine the next iteration step in the designof the SC gantry.

——————————————–

ADJOINT VARIABLE METHOD FOR RAPIDDESIGN OPTIMIZATION OF ELECTROSTATIC

LENS SYSTEMS

L. T. Neustock, P. C. Hansen, Z. E. Russell, L. Hesselink

Stanford University, Ion Innovations

Keywords: optimization, electrostatic lens design, finite elementmethod, inverse design, electrostatic field solver, adjoint variablemethod, charged particle optics, einzel lens, velocity verlet

We have implemented a generative algorithm for design opti-mization of electrostatic charged particle optical devices usingthe discrete adjoint variable method. In this work, we optimizea series of electrostatic lenses to minimize spherical and chro-matic aberrations and perform beam steering. To the best ourknowledge, these are the first charged particle optics systems de-signed by an adjoint variable based algorithm. Physical systemswith many designable parameters (e.g. dimension, shape and ap-plied voltages) are computationally burdensome to optimize. Toefficiently improve the device design it is crucial to know howits performance changes under all designable perturbations to itsshape, dimensions, and operating conditions. This sensitivity ofthe design to its design parameters can be obtained by simulatingeach potential perturbed device in turn. This is a computationallycostly approach requiring at least one extra full-system calcula-tion per design parameter. In contrast, adjoint design sensitivityanalysis is a method to obtain sensitivities to all design param-eters at once through an algorithm with nearly-fixed computa-tional cost. Thus, adjoint methods enable rapid optimization ofcomplex systems. This has led to the pervasive use of these meth-ods in aeronautical, structural and photonic design. To obtainsuch a rapid algorithm for charged particle optics, we derived andimplemented a fully discrete adjoint system solver for the non-linear, coupled system comprising the Laplace equation for theelectric potential and an equation of motion based on the Lorentzforce law and Newton’s second law, using a custom-built elec-trostatic finite-element method and a charged particle dynamicssimulator based on Verlet integration. This solver allows for arbi-trary selection of initial designs and number of design parameters

and calculates the sensitivity of the design to charged particle tra-jectories for each of these parameters. The method demonstratedcan be applied to miniaturizing complex systems and optimizingmulti-beam applications such as lithography.

——————————————–

INVESTIGATION OF ELECTROMAGNETIC-SYLCFOR CHROMATIC ABERRATION CORRECTION

R. Nishi(1), S. Hoque(1, 3), H. Ito(1,2) and A. Takaoka(1)

(1) Osaka University, (2) Hitachi High-TechnologiesCorp., (3) Hitachi High-Technologies America, Inc.

Keywords: chromatic aberration, symmetric line currents, SEM,aberration correction

We are studying a simple aberration corrector for scanning elec-tron microscopes (SEM). We have proposed SYLC (SYmmetricLine Currents) corrector in which parallel line currents are sym-metrically arranged instead of usual magnetic multipoles to cor-rect spherical aberration [1, 2, 3]. The main feature of SYLCis being free of magnetic material, thus eliminating the prob-lems of hysteresis, nonuniformity and magnetic saturation of themagnetic material. In addition to spherical aberration, correc-tion of chromatic aberration is also necessary for low accelera-tion voltage SEM. Therefore, SYLC is expanded to incorporateelectrostatic multipole field with magnetic multipole field. Ap-plying positive and negative electrostatic potentials alternately tothe conducting lines of SYLC yields a superposition of electro-static and magnetic 2N-poles, which we call an electromagnetic-SYLC. Moreover, since the direction and distribution of 2N-poleelectric and magnetic field are the theoretically same, the numberof poles can be reduced by half comparing with a conventionalelectromagnetic multipole. We show that chromatic aberrationcan be corrected by a model in which the conventional multipolesof the quadrupole 4-stage corrector proposed by H. Rose [4] arereplaced by the combination of electro-and magnetic-SYLC. [1]Nishi R., Ito H., Hoque S., (IMC2014), IT-1-P2984, pp.200-201[2] S. Hoque, H. Ito, R. Nishi, A. Takaoka, E. Munro, Ultra-microscopy 161, (2016) 74-82 [3] P. W. Hawkes and E. Kasper,Principle of Electron Optics, Vol.2 : Applied Geometrical Op-tics, 2nd ed., chap. 41, pp. 986-988, Academic Press, 2017. [4]H. Rose, Optik 32, (1970) 144

——————————————–

EVALUATION OF A MONOCHROMATOR WITHOFFSET CYLINDRICAL LENSES FOR

ELECTRON MICROSCOPY

Takashi Ogawa

Korea Research Institute of Standards and Science

Keywords: Monochromator, Cylindrical lens, Electronmicroscope

Monochromators (MCs) have been indispensable optical compo-nents for advanced electron microscopes [1-3]. MCs can improveimage resolutions at low energy conditions and energy resolu-tions of EELS spectra. Both are achieved by narrowed energyspreads of electron beams. At the CPO9 conference, a new MC

14

with cylindrical lenses (CLs) was proposed [4]. The MC con-sists of two CLs in offset layout with the middle plane symme-try. The strongly excited CLs generate large energy dispersions,which enables the MC optics with high performance and simplestructure. Based on theoretical studies on the MC [5-7], a proto-type of the MC has been constructed at high mechanical accuracyand combined with highly stable electronics. The MC achievedthe energy resolution of 73 meV by measuring energy distribu-tions with an additional energy analyzer [8]. Observation in twoconditions with or without the MC confirmed fine beam profilesbecause of the symmetry of the MC. The results assure applica-bility of the MC to electron microscopes. At the conference, re-cent evaluation results of the MC will be presented. References:[1] H. Rose, Ultramicroscopy 78 (1999) 13. [2] H.M. Mook, P.Kruit, Ultramicroscopy 81 (2000) 129. [3] O.L. Krivanek, et al.,Nature 514 (2014) 209. [4] T. Ogawa, et al., Microsc. Microanal.21 (S4) (2015) 112, Proceedings of CPO9. [5, 6] T. Ogawa, B.Cho, Nucl. Instrum. Methods. A 772 (2015) 5, 800 (2015) 18.[7] T. Ogawa, et al., J. Vac. Sci. Technol. B 33 (6) (2015)06FJ01-1-11. [8] T. Ogawa, Y. Takai, J. Vac. Sci. Technol. B 36(3) (2018) 032902.

——————————————–

SIMULATION OF MOTION OF MANY IONS IN ALINEAR PAUL TRAP

Martin Oral

Institute of Scientific Instruments of the CAS,Kralovopolska 147, 612 64 Brno, Czech Republic

Keywords: linear quadrupole Paul trap, Coulomb crystals,optical atomic clock, micromotion, trajectory simulation

The quadrupole linear Paul trap is one of the key instruments inbuilding highly stable atomic clocks. However, a frequency ref-erence based on a single trapped ion is limited in stability due tothe time needed for the interrogation cycle which cannot be fur-ther shortened. A promising strategy is the utilization of multipletrapped ions. The ions of the same kind then repulse each otherwith the Coulomb force, which is countered by the ponderomo-tive force of the time depended field in the trap. A few ions forma chain along the axis of a linear Paul trap. Adding more ions (afew tens or hundreds) gives rise to Coulomb crystals. We createdan efficient simulation code which calculates the motion of suchcollections of ions in quasistatic radiofrequency fields of real lin-ear quadrupole traps (including the micromotion). We attempt totake into account various methods of cooling the ions. The sim-ulation tool can be used to study the formation and the dynamicsof Coulomb crystals under conditions corresponding to variousexperimental set-ups.

——————————————–

SIMULATION FOR THE DEVELOPMENT OFPRECISE AUTO-FOCUSING OF SEM LENSES

Yoichi Ose and Makoto Ezumi

Hitachi High-Technologies Corp.

Keywords: SEM, auto-focus, self-inductance, aftereffect

The latest SEMs in industrial product lines should meet the de-mands of higher resolution and higher throughput. The higherresolution objective lenses have the shorter depth of focus whichrequire the more precise adjustment of their coil current. Theduration of auto focusing should be shorter to obtain higherthroughput then the precision of auto-focusing is subject to thedelay of current response with the self-inductance and the after-effect of magnetization. We propose a response function basedon those two delay mechanisms while the ratio of the aftereffectis estimated comparing experimental delay times with simulatedones. A function to calculate self-inductance has been built ontoan in-house electron optics simulator “EMB2D” which shows thesaturation of self-inductance over the permeability of 1000. Wedemonstrate the ratio of the aftereffect 0.06 with time constantof 1.0 sec. is plausible in this case and our proposed responsefunction is useful to improve the precision of auto-focusing.

——————————————–

WIDE-ANGLE ANNULAR ELECTRON BEAMFOCUSING COLUMN.

Balamuniappan Pranesh, Ang Wei Kean, AnjamKhursheed

National University of Singapore

Keywords: annular focused electron/ion beams, aberrationcorrection, electric sectors

There are many applications in electron microscopy, electronspectroscopy, as well as accelerator physics that require the com-bination of minimizing a focused electron beam’s probe size,while maximizing its beam current. This paper describes howit can be done through the use of annular focused electron beamcolumn designs and cold field emission sources, where an elec-tron beam is propagated and focused in the form of a ring beam.For relatively small probe semi-angles, where the central anglelies between 1◦ to 2◦, a 3 stage deflector-corrector in combina-tion with an objective lens will be presented. The objective lensdesign can be either magnetic or electric. For probe semi-anglesabove say 20◦, a column consisting of 2 identical electric sectorsand 2 identical focusing lenses functioning with odd symmetrywill be presented. The column is designed to cancel energy dis-persion while limiting geometric aberrations to be of 3rd orderat the point of final focus. Both designs predict to have over twoorders of magnitude higher beam current than their correspond-ing conventional electron beam focused columns for the samefinal probe size. For a 1.2 keV annular electron beam with asemi-angular spread of 45◦ +- 0.1146◦ semi-angle, simulationresults predict a final spot size of 5.7 nm at a working distance of3.4 mm, giving approximately three orders of magnitude largerprobe current than its corresponding conventional on-axis elec-tron beam column for a similar probe size.

15

——————————————–

COMPARISON OF BEAM OPTICS FORNORMAL-CONDUCTING AND

SUPERCONDUCTING GANTRY BEAMLINEAPPLIED TO THE PROTON THERAPY SYSTEM

Bin Qin

Huazhong University of Science and Technology

Keywords: Proton therapy, beam optics, superconducting gantrybeamline, high order effect

Due to the unique ‘Bragg peak’ dose distribution characteristicsof the proton beam, the proton therapy is recognized as one ofthe most precise and effective radiotherapy method for tumors.A gantry is required to project the beam on tumors with variousangle for multiple fields radiation, and a superconducting beam-line can significantly reduce the size and the weight of the gantry.A proton therapy system is under development in HUST. Thispaper will introduce the comparison study of the beam optics fornormal-conducting and superconducting gantry beamline. Beamsimulation study which demonstrates the influence of high ordermagnetic field effect in the beamline, will also be described.

——————————————–

DIFFERENTIAL ALGEBRAIC METHOD INELECTRON OPTICAL DESIGN

Tomas Radlicka

Institute of Scientific Instruments CAS

Invited TalkKeywords: differential algbraic method, correctors, aberrations,parasitic aberrations

The differential algebraic method is standardly used in parti-cle accelerator design. It provides the high order transfer mapneeded for simulations of the long-term ring stability. On theother hand, the electron optics community prefers methods hav-ing a solution in the form of the aberration integral which pro-vides a deeper view of the effect of each element on the electronoptical properties of the system but they became too complicatedwith increasing aberration order, which is not a case of the differ-ential algebraic method. We show the potential of the differentialalgebraic method on a design of electron optical systems. Thefirst part deals with simulation of the axially symmetrical sys-tems where the effect of the fringing field regions must be han-dled to obtain correct values of aberration coefficients. The nextpart shows the application of the method on aberration correc-tor design, we propose the efficient combination of the trajectorymethod used for a qualitative description of the primary aberra-tion behavior and the differential algebraic method providing thequantitative description of electron optical properties. Instead ofthe aberration integrals the shape of the aberration coefficients isused for estimation of the effect of each electron optics element.The approach is used in the analysis of parasitic aberrations andoptimization of the system with corrector. The last part showsthe application on the systems whose symmetry was perturbatedby the general 3D elements of the in-lens detector.

——————————————–

RECENT SUCCESSES OF MULTI-REFLECTIONDEVICES AT RIKEN’S RIBF FACILITY AND

SOME THOUGHTS ABOUT HIGHLY ACCURATEMASS CALIBRATION USING ION TRAPS

M. Rosenbusch (1), Y. Ito (2), P. Schury (3), M. Wada (3),H. Wollnik (4), and the SHE-Mass Collaboration

(1) RIKEN Nishina Center (RNC), (2) Japan AtomicEnergy Agency (JAEA), (3) KEK - Wako Nuclear Science

Center (WNSC), (4) New Mexico State University(NMSU)

Keywords: Multi-Reflection Devices, Precision MassSpectrometry, Mass Calibration, Nuclear Masses

By the use of coaxial ion mirrors, very long flight paths of ions toa time-of-flight detector were possible since the early 90’s, whenmulti-reflection time-of-flight mass spectrometry (MRTOF MS)was invented [1]. The development and usage of MRTOF massspectrographs has been performed intensely at nuclear-physicson-line laboratories at later time and is still being continued. Dueto the high precision of ion masses achieved in times <10 ms,and the single-ion sensitivity achievable for TOF MS, this tech-nique became attractive for high-precision mass determinationsof rare and short-lived nuclei. Due to minute production ratesand insufficient beam emittances provided by the existing facil-ities, the MRTOF technique has been developed in combinationwith ion-trapping techniques, i.e. quadrupole ion traps as ac-cumulators, ion coolers, and injectors into the MRTOF device.At the super-heavy element (SHE)-mass facility of RIKEN-KEK[2] the exotic isotopes of 249-253Md [3], many other rare specieslike 210-214Ac/Ra [4] and about 70 other isotopes have success-fully been mass determined with precisions down to several hun-dred ppb [5]. Reaching extremes for mass precisions, new con-siderations of the effect of ion traps on the mass calibration ofan MRTOF mass spectrograph may become important and havebeen calculated. Here, the success of a MRTOF MS at RIKEN’sRIBF facility will be presented and later-on, some thoughts onmass calibrations for precisions beyond the so far achieved oneswill be discussed. [1] H. Wollnik and M. Przewloka, IJMS IonProc. 96, 267 (1990) [2] P. Schury et al., Nucl. Instr. Meth.B 335, 39 (2014) [3] Y. Ito et al., Phys. Rev. Lett. 120, 102501(2018) [4] M. Rosenbusch et al., Phys. Rev. C 97, 064306 (2018)[5] S. Kimura et al., IJMS 430, 134 (2018)

——————————————–

16

AN ALGORITHM FOR CHARACTERIZING THEGEOMETRIC OPTICS OF CHARGED PARTICLE

INSTRUMENTS

Khashayar Shadman, Marian Mankos

Electron Optica, 1000 Elwell Court #110, Palo Alto,California 94303, USA

Keywords: nonlinear ordinary differential equation solver,ray-optical simulator

The proper design of charged particle instruments requires ac-curate simulations of the particle trajectories within the appliedelectromagnetic fields. Of interest, in particular, are the smalldeviations in the trajectories from the optical axis of the instru-ment. These deviations have linear and nonlinear dependencieson the initial conditions of the particles whose coefficients de-fine the optical properties of the instrument. These coefficientscan be computed directly, circumventing the need to simulate theindividual trajectories. The traditional methods apply perturba-tion theory to the equation of motion to manually derive integralexpressions for the primary nonlinear (aberration) coefficients.However, the manual procedure is difficult to extend for the com-putation of the higher order aberrations as the algebra becomesintractable. This difficulty has been overcome by the differen-tial algebraic method, which uses a structure from nonstandardanalysis to compute the series expansions of the electromagneticforces about the optical axis in an automated manner. Here, analgorithm is presented that uses a standard mathematical tech-nique, the binomial theorem, to codify this calculation. This al-gorithm gives the evolution of the coordinates along the opticalaxis as a series expansion in their initial values. The series co-efficients can be derived up to any order in one of two ways, byeither accumulating them serially along the optical axis via thesolution to a difference equation or in parallel over the entire re-gion of interest via an iterative solution of an integral equation.The algorithm is not limited to the equation of motion. It can beapplied to higher order nonlinear, ordinary differential equationsfor the evolution of the dependent variables in the neighborhoodof a principal path.

——————————————–

A HIGH RESOLUTION MULTI-TURN TOF MASSANALYZER

V. Shchepunov, M. Rignall, R. Giles, R. Fujita, H.Nakanishi, and H. Waki

Shimadzu Research Laboratory (Europe) Ltd, Manchester,M17 1GP, United Kingdom

Keywords: Mass analyzer, TOF mass spectrometer, Ion optics,Aberrations

Ion optical design of a high resolution Multi-turn TOF Mass An-alyzer (MT-TOF MA) is presented. The analyzer has rotationalsymmetry of the main electrodes, which allows higher densityof turns in the azimuthal (drift) direction compared to MT-TOFMA’s linearly extended in the drift direction. The analyzer ge-ometry has mid-plane symmetry and comprises a pair of polar-toroidal sectors S1 (lower) and S3 (upper), a toroidal sector S2located at the mid-plane of the system, a pair of polar (trans-axial) lenses, and a pair of conical lenses for longitudinal andlateral focusing, each pair of the electrodes being mirror sym-metric with respect to the mid-plane. Additionally, drift focusingsegments embedded into S2 electrodes are used to provide focus-ing and spatial isochronicity in the drift direction. Due to an openreference trajectory and static electric fields the analyzer retainsthe full mass range of the injected ions. Geometry and poten-tials of the analyzer electrodes are optimized for 5-8 keV ions toprovide transverse focusing and isochronicity with some higherorder corrections of TOF aberrations. Ion optical properties ofthe analyzer are described in detail. Several operational modesof the analyzer are feasible. At small turn numbers (up to∼10-12turns) focusing in the drift direction is not required, and the turnnumber can be varied by simple injection steering. Maximumm/dm in this mode is ∼50-60 k (fwhm). At larger turn numbersthe drift focusing must be used. The ions’ drift in the azimuthaldirection can be reversed by using a pair or dedicated reversingdeflectors. This gives possibility of multiple passes in the driftdirection. It was demonstrated earlier that ∼200 k (fwhm) ofmass resolving power is feasible after 2 passes in the drift di-rection (forward and reversed). Further optimization of the ana-lyzer for 4 passes allowed us to increase m/dm up to∼400-500 k(fwhm). Respective numerical simulations are presented. Apartfrom TOF operational modes the analyzer can be optimized forthe use as an electrostatic trap with Fast Fourier Transform (FFT)mass analysis.

——————————————–

FFT AND TOF OPERATIONAL MODES OF AHYBRID MASS ANALYZER

V. Shchepunov, M. Rignall, R. Giles, R. Fujita, H.Nakanishi, and H. Waki

Shimadzu Research Laboratory (Europe) Ltd, Manchester,M17 1GP, United Kingdom

Keywords: FFT mass spectrometry, Mass analyzer, Ion optics,Aberrations

It was demonstrated earlier that a rotationally symmetric massanalyzer can achieve mass resolving power of∼200 k (fwhm) ina multi-turn TOF operational mode. The analyzer geometry has

17

mid-plane symmetry and comprises a toroidal sector S2 locatedat the mid-plane, a pair of polar-toroidal sectors S1 and S3, pairsof trans-axial and conical lenses for lateral and longitudinal fo-cusing, each pair being mirror symmetric with respect to the mid-plane. In the multi-turn TOF operational mode drift focusingsegments are used to provide focusing and spatial isochronicityin the drift direction. In this work we present results of simula-tion studies demonstrating that analyzers with similar geometriescan be used as (i) a pure Fast Fourier Transform (FFT) mass an-alyzer with image charge detection providing m/dm of at least∼800 k (fwhm), (ii) a pure multi-turn TOF mass analyzer withm/dm of at least ∼200 k (fwhm), and (iii) a hybrid instrumentproviding either m/dm ∼100 k (fwhm) in multi-turn TOF modeor m/dm ∼800 k (fwhm) in FFT mode. Analyzers of three dif-ferent sizes (500 mm, 250 mm and 120 mm of external diameterof S2) have been studied numerically for ions at 5-8 keV energy.The largest analyzer is the best for the use in the multi-turn TOFmode. Its simulated m/dm for 400 Da ions at 5 keV is ∼200 k(fwhm) at typical flight times of about 1.1 ms. Large size makesthe analyzer rather slow for running it in FFT mode. On the con-trary, the smallest analyzer is the fastest of the three and the mostappropriate for the use in the FFT mode. The 5th harmonic of theFFT signal provides m/dm of∼800 k (fwhm) after 1 sec of mea-surement time. Its estimated m/dm in the multi-turn TOF mode isonly∼15-20 k. Hybrid anayzer of the intermediate size (250 mmof S2 diameter) demonstrates m/dm ∼100k (fwhm) in the multi-turn TOF mode or m/dm of ∼800 k (fwhm) at 2.1 s measure-ment time in the FFT mode. Similar instruments can be run inone of the two complimentary modes - the multi-turn TOF modewith lower m/dm and faster mass analysis, or the FFT mode withhigher m/dm and slower mass analysis. Corporate software andSIMION program were used in the studies.

——————————————–

ROBUSTNESS CALCULATION OF MAGNETICSECTORS USING DIFFERENTIAL ALGEBRAIC

METHOD

Yasuhiro Shirasaki and Momoyo Enyama

Hitachi, Ltd. Research and Development Group

Keywords: SEM, magnetic sector, simulation

In electron microscopes, beam separators are important electronoptical components used to separate the signal electrons. In par-ticular, magnetic sectors have a long history of being used for lowenergy electron microscopes, and in recent years have also beenconsidered for use in scanning electron microscopes (SEMs). Tobe able to use magnetic sectors, the aberrations introduced mustbe negligible compared to that introduced by other optical com-ponents [1]. Furthermore, the magnetic sectors must be robustenough so that they can be fabricated and operated under practi-cal machining and current source precisions, respectively. Aber-rations introduced by magnetic sectors can be calculated usingthe differential algebraic method to solve the electron beam tra-jectory as a function of its initial conditions, which are position,direction, and energy [2]. However, to assess their robustness theaberrations must be calculated for many cases, each with differ-ent pole piece shapes and excitation currents caused by the ma-chining error and the current source noise. This becomes tediousand increasingly difficult as the number of pole pieces becomes

large. To make the assessment of robustness possible, we havedeveloped a simulator that treats the angles of the grooves form-ing the pole pieces and the excitation currents of the pole piecesas parameters that are part of the initial condition. Therefore, inaddition to the initial conditions of the beam, the aberrations canbe calculated also as a function of the groove angles and the ex-citation currents. Using this simulator, we evaluated the robust-ness of a chicane-type magnetic sector [3], which only deflectsthe secondary beam, and found that this chicane-type magneticsector can be useful for electron optical systems with straightoptical axes such as SEMs. References: [1] Y. Shirasaki et al.,EIPBN 2016 (2016). [2] M. Berz, Modern Map Methods in Par-ticle Beam Physics. Academic Press (1999). [3] V. Kolarik et.al., “Close packed prism arrays for electron microscopy”, Optik87, 1 (1991).

——————————————–

PRISMATIC MASS SPECTROGRAPH WITH ACONICAL ACHROMATIC PRISM AND

TRANSAXIAL LENSES

A.A. Sapargaliev, I.F. Spivak-Lavrov, A.A. Trubitsyn,E.M. Yakushev

Kazakhstan, Russia

Keywords: prismatic mass-spectrometer, transaxial lens

The conical achromatic prism (CAP) has a record angular disper-sion equal to about 50 radians per 100% of mass variation [1]. InCAP, electric and magnetic fields are realized whose potentials ina spherical coordinate system depend only on angular variables.The particles of a homogeneous planar parallel ion beam move inthe middle plane of the CAP along similar trajectories and main-tain parallelism at the exit from the CAP. The CAP also focuseson energy, and the parallelism of the volume beam is maintaineddue to its telescoping in the vertical direction. CAP can be usedin a prismatic mass spectrometer, which in its scheme is similarto a prism light-optical spectrometer equipped with a collimatingand focusing lens. The linear dispersion of the prism spectrom-eter is equal to the angular dispersion of the CAP multiplied bythe focal length of the focusing lens. A prismatic device is de-signed in which three-electrode transaxial lenses are used as acollimating and focusing lens. Due to the large mass dispersionby using a positional detector located in the focal plane of thefocusing lens, a mass spectrograph can also be implemented insuch a device. 1. Spivak-Lavrov I.F. Analytical Methods forThe Calculation and Simulation of New Schemes of Static andTime-of-Flight Mass Spectrometers. - Advances in Imaging andElectron Physics. - Vol. 193, Burlington: Academic Press, 2016.- P. 45-128.

——————————————–

18

DRIFT CONTROL IN ISOCHRONOUSMULTI-REFLECTION TOF ANALYZER WITH

ELONGATED ION MIRRORS

Dmitry Grinfeld, Christian Hock, Hamish Stewart,Alexander Makarov

Thermo Fisher Scientific

Keywords: mass spectrometry, aberration correction, ion mirror

Ion mirrors are employed to establish a many meters-long flightpath in a table-size time-of-flight (ToF) mass spectrometer [1-3]to enhance the mass resolving power. This goal also requiresisochronicity, i.e., independence of the travel time with respectto the ion’s initial coordinates and velocities. The systems underconsideration comprise a pair of ion mirrors facing each otherin such a manner that the ions isochronously oscillate betweenthem. To achieve spatial separation between the oscillations, themirrors are elongated in the direction orthogonal to the line of os-cillations. The ions are injected at an angle and slowly drift in thedirection of elongation describing zigzag paths. An unavoidablespread of the drift velocities results, however, in the expansionof the ion bunch, and adjacent oscillations start overlapping afterjust a few reflections. To counteract the drift expansion, the ideal’isochronous 2D electrostatic field f0(z,x) should be given a rela-tively small perturbation df(z,x,y) that varies in the drift coordi-nate ‘y’. Some designs [3] incorporate an array of three-electrodelenses placed after each reflection, while in other solutions [4] themirrors are sectioned to reverse the drift direction and refocus theion bunch. It should be noticed, however, that the y-dependentfield breaks the reflection isochronism so that the time of flightbecomes a function of the ion’s injection position y(0) and theinjection angle dy/dz, which should be avoided. In this presenta-tion, we develop a perturbation theory of drift control and its im-pact on the oscillation isochronism. It is shown that some classesof field perturbations generated by inter-mirror electrodes andslight non-parallelism of the ion mirrors counteract the drift ex-pansion while the isochronicity is practically retained. As sim-ulations have demonstrated, these results may be used to buildToF mass spectrometers with improved mass resolving powerand throughput. [1] Nazarenko L.M., Sekunova L.M., Yaku-shev E.M., USSR Patent SU1725289 (1989) [2] Wollnik H. andCasares A., Int. J. Mass Spectrom. 227 (2) (2003): 217-22 [3]Yavor M., Verentchikov A., et al., Physics Procedia 1 (2008):391-400 [4] Sudakov M. and Kumashiro S., Nucl. Instr. MethPhys Res. A645 (2011): 210-5

——————————————–

STATISTICAL COULOMB INTERACTIONS INMULTI-BEAM SEM

Jan Stopka, Pieter Kruit

Institute of Scientific Instruments of the CAS, v. v. i.;Faculty of Applied Sciences, Delft University of

Technology

Keywords: Multi-beam; Coulomb interactions; Trajectorydisplacement

Statistical Coulomb interactions in conventional scanning elec-tron microscopy mostly affect the probe size via energy spread

and virtual source broadening in the emitter vicinity. However,in a multi-beam probe forming system such as multi-beam SEM,the trajectory displacement due to interactions in the whole col-umn can give a major contribution to the final probe size. A the-oretical description of trajectory displacement is only known forsingle-beam systems. It can be expressed using approximate an-alytical formulae for the total trajectory displacement in a beamsegment (Jansen’s theory) or by integrating contributions of in-finitesimally thin beam slices (the slice method). We build onJansen’s theory of statistical Coulomb interactions and developanalytical formulae for the trajectory displacement in a multi-beam system. We also develop a more precise semi-analyticalresult using the slice method. We compare both approaches witha Monte-Carlo simulation and show a good agreement thereof.Finally, we discuss the implications of our results for the opti-cal design of multi-beam SEM. In a multi-beam with probe sizedominated by Coulomb interactions, an increase in the number ofbeamlets does not necessarily provide an increase of throughput,because the probe size is limited by the total current. Further-more, we disprove the notion of “the fewer crossovers - the lesscoulomb interactions” by showing the quadratic dependence oftrajectory displacement on segment length.

——————————————–

ACCURATE MEASUREMENT AND CORRECTIONOF IMAGE DISTORTION IN TEM WITHOUT

REFERENCE

Hirokazu Tamaki 1, Yoichi Hirayama 2, Hiromi Inada 2

1 Research & Development Group, Hitachi, Ltd., 2Science & Medical Systems Business Group, Hitachi

High Technologies Corporation

Keywords: Image Distortion, Aberration, Electron Optics

As well known, transmission electron microscope (TEM) has theimage distortion due to lens aberration, and it is not easy to re-move. Image distortion lowers the accuracy of image scale andincreases the error in computational image reconstruction such asTomography, Single Particle Analysis, and Ptychography. Oneof the solutions to this problem is a distortion correction with aknown information. However it is difficult to get an exact infor-mation of the distortion, because there are no suitable measure-ment references for high magnification in electron microscopes.To solve this problem, we developed a new method to quanti-tatively measure the image distortion without reference. Thismethod uses multiple images which are taken in the differentfields of view with some amount of overlap. Between each im-age, there are mainly two kinds of changes. One is a linear im-age shift which corresponds to the field of view movement andthe other is a nonlinear image transformation which originatesfrom the image distortion. From the nonlinear image transfor-mations, differential component of distortion can be obtained,which is then used to calculate image distortion. From numer-ically generated test images, measurement error less than 0.1%was confirmed for radial distortion which is the main cause ofbarrel or pincushion distortion. Moreover, post-correction of theimage distortion from the measured results and the applicabilityto experimental images were also confirmed to be possible. Thismethod allows many kinds of typical specimen to be used for dis-

19

tortion measurement, and the result can be used for both opticalsystem evaluation and real-time or post distortion correction.

——————————————–

MUON LOSSES FROM BETATRON RESONANCESAT THE MUON G-2 EXPERIMENT AT FERMILAB

David Tarazona, Martin Berz, Kyoko Makino

Michigan State University

Keywords: betatron resonance muon losses

The Muon g-2 Experiment (g-2) at Fermilab is directed towardmeasuring the muon anomalous magnetic moment with statisti-cal and systematic relative errors smaller than 140ppm. This newmeasurement will serve as strong indication of yet undiscoveredparticles beyond the Standard Model and validate or disproveother theoretical models beyond the SM. Of special interest isthe reduction of muon losses to achieve the precision needed atthe g-2 Experiment. For this purpose, we have developed a de-tailed and precise g-2 Storage Ring using COSY INFINITY thatconsider inhomogeneities of the magnetic field, up to the elec-tric 20-pole multipole of the Electrostatic Quadrupoles (ESQ),and injection to the ring based on measurements. In specific, wehave recreated lost muons rates for several possible configura-tions of the ESQ system in order to find the best possible scenar-ios that minimize muon losses. Additionally, comparison withmeasurements have allowed to identify possible sources of er-ror due to both the beam dynamics and imperfections in the g-2storage ring.

——————————————–

OPTICAL DESIGN OF THE FIXED FIELDPERMANENT MAGNET GANTRY FOR THE

PROTON CANCER THERAPY

D. Trbojevic 1, William Lou 2, Stephen Brooks 1,Francois Meot 1, and Nicholaos Tsoupas 1

1. Brookhaven National Laboratory 2. Cornell University

Keywords: Fixed Field magnets, large momentum acceptance,hadron cancer therapy, proton gantry

We present an optical design of the proton therapy gantry with avery large momentum acceptance of dp/p=+-33%. This momen-tum range corresponds to the kinetic energy of protons withina range of 65-250 MeV - energies required for the patient pro-ton radiation therapy moving the Bragg peak between 3.5-38cm. The optics uses combined function magnets with a fixedlinear magnetic field. The permanent magnets of the Halbachtype, are made of Neodymium Iron Boron - NeFeB. Additionalimportant optics properties is that the gantry is made of chro-matically matched parts. The first achromatic part bends protonsupwards. The central momentum particles travel in a perfect cir-cle while the rest of them radially oscillate with orbit offsets of+14 -10 mm; positive offsets correspond to the higher while thelower energies, accordingly. The same magnets from the halfof the first gantry part continue symmetrically upward but withopposite bending direction. The last part of the gantry bringsfocused proton beams to the patient. The last gantry magnet is

more than 1.2 meters away from the patient. Two radially scan-ning magnets are placed within this space to allow +/- 10 cm inthe horizontal plane at the patient. There are multiple advantagesof this optical design: the large momentum acceptance allowsfast energy change without magnetic field variation. The onlyvariable magnetic field comes from the scanning magnets at theend of the gantry. The transverse scanning is significantly sloweras the longitudinal energy scanning occurs for each radial po-sition. The patient treatment time is shorter and the operationis simplified due to the fixed magnetic field. There is a signifi-cant reduction of power consumption as well as with the overallgantry cost; it is one order of magnitude reduced with respect tothe other existing gantries. The weight and size of the magnets issignificantly reduced allowing lighter rotating structures.

——————————————–

THE ERHIC SPIN ROTATOR AND THE BEAMOPTICS OF THE 400 MEV TRANSFER LINE TO

RCS

N. Tsoupas*, M. Blaskiewicz, H. Lovelace III, F. Meot,C. Montag, V. Ptitsyn, V. Ranjbar, S. Tepikian, W. Zhang,

G.M. Wang, E. Wang, W. Weng, F. Willeke

Brookhaven National Laboratory

Keywords: Spin Rotator

The 400 MeV LINAC [1] is the first acceleration stage of theelectron accelerator of the proposed eRHIC collider [1]. Thesecond acceleration stage of the electron bunches is the RapidCycled Synchrotron (RCS) which can increase the energy of theelectron bunches up to 18 GeV. The function of the transfer linebetween the LINAC and the RCS (LtRCS) is twofold, first totransfer the electron beam from the exit of the 400 MeV LINACto the injection point of the Rapid-Cycled-Synchrotron (RCS)[1] and second to rotate the electron spin from the longitudinaldirection to the vertical. We will describe the beam optics ofthe transfer line, and the required constrains on the beam linefor the proper beam matching of the 400 MeV LINAC and theRCS and the spin rotation. A section will also be devoted to dis-cuss the spin rotator. * tsoupas@bnl.gov [1] PCDR of eRHICBrookhaven National Laboratory

——————————————–

DEVELOPMENT OF A DIAGNOSTIC SETUP FORQUANTUM ELECTRON MICROSCOPY

M. Turchetti, N. Abedzadeh, A. Agarwal and K. K.Berggren

Massachusetts Institute of Technology

Keywords: quantum electron microscopy, aberration,ptychography, shadow imaging, ronchigrams

Quantum electron microscopy is one of the most promising ap-proaches that could overcome the resolution limit imposed bythe radiation damage especially to biological samples. This mi-croscopy scheme requires the design of novel components suchas gated electron mirrors, and the development of a platform fordiagnostics of ultra-fast electron optics. This testbed would allow

20

time-resolved characterization of such elements either isolated orintegrated into the system, both in transmission and in reflectionmodes. In this work, we propose a diagnostic setup comprisedof a ptychography and shadow-imaging (Ronchigrams) platformto allow the evaluation of wavefront aberrations in transmission,and a procedure for beam characterization in reflection, whichemploys an electron mirror with tuneable spherical aberration.

——————————————–

ANALYTIC ABERRATION FORMULAS ANDTRANSFER MAPS OF ELECTROSTATIC

DEFLECTORS

Eremey Valetov and Martin Berz

Michigan State University

Keywords: electrostatic deflectors, transfer maps, aberrations,differential algebra

Using an iterative perturbation method, we derived first and sec-ond order analytic aberration formulas for the electrostatic de-flector, specified by the curvature radius, central angle, and inho-mogeneity coefficients. We compared the results with those ofnumerical differential-algebraic (DA) integration of the ODEs ofmotion using COSY INFINITY. Additionally, we directly calcu-lated the transfer map of the electrostatic spherical and cylindri-cal deflectors in the laboratory coordinate system using a Runge-Kutta integrator. For the electrostatic spherical deflector, we alsocalculated the transfer map analytically and in closed form us-ing the properties of the respective elliptical orbits from Keplertheory. We compared the results with the DA transfer map ofCOSY INFINITY’s built-in electrostatic and spherical deflectorelements, as well as with the program GIOS.

——————————————–

TRANSVERSE PHASE SPACE TOMOGRAPHY INBEAMLINES

Adam Watts, Carol Johnstone

Fermilab

Keywords: optics, tomography, transverse, phase space

Methods of reconstructing beam transverse phase space usingcomputed tomography are compared and optimized in simula-tion to improve reconstruction accuracy. Errors and artifacts areshown for both Filtered Back Projection (FPB) and Simultane-ous Algebraic Reconstruction Technique (SART) methods as afunction of their respective free parameters. Finally, a theoreti-cal discussion of common optics configurations and their effecton the feasibility of computed tomography is shared, and adviceis given on choosing optimal sections of rings or beamlines toperform such reconstructions.

——————————————–

A TRANSVERSE ENVELOPE MACROPARTICLEMETHOD FOR MODELING HIGH-GAIN FREE

ELECTRON LASERS

Stephen D Webb

RadiaSoft, LLC

Keywords: free-electron lasers, reduced models

Recirculating schemes for free-electron lasers, such as the re-generative amplifier FEL, require simulating thousands of passesthrough a high-gain FEL system. Modeling high-gain FELs ac-curately requires the inclusion of diffractive effects in the radi-ation field and transverse motion of the electron beam. Conse-quently, most simulation codes that study high-gain FELs arethree-dimensional self-consistent codes. Using these codes inthe aforementioned multi-pass system would be too computa-tionally expensive to produce results in a reasonable amount oftime. However, because the transverse dynamics for the electronbunch is largely comprised of linear motion through drifts andquadrupoles, it can be represented by carefully constructed re-duced models. Here we present the use of macroparticles with atransverse gaussian envelope to map the three-dimensional prob-lem to a quasi-one-dimensional problem, reducing the compu-tational cost of simulating the high gain FEL. We present thederivation of the algorithm and preliminary benchmarking sim-ulations against GENESIS for both accuracy and computationalefficiency.

——————————————–

COMPUTATION AND CONSEQUENCES OFHIGH-ORDER AMPLITUDE AND MOMENTUMDEPENDENT TUNE SHIFTS IN THE MUON G-2

RING

Adrian Weisskopf, David Tarazona, Martin Berz

Michigan State University

Keywords: Betatron tune shifts, moun g-2 ring, normal form,high-order transfer maps

Betatron tune shifts can influence the coherent betatron oscilla-tion (CBO) frequency of a mismatched beam in an accelerator,in particular the muon beam in the storage ring of the Muon g-2Experiment at Fermilab (E989). In this case, nonlinear electricfields from the electrostatic quadrupole system (EQS) installedwithin the storage ring to confine muons vertically and other non-linearities due to slight errors in the uniformity of the magneticfield produce substantial amplitude-dependent tune shifts. In ad-dition to this, the storage ring momentum acceptance of +/-0.5allows for momentum-dependent tune shifts. Motivated by theseaspects and by the sensitivity of the final measurement precisionat E989 to the CBO frequency, we present a normal-form basedmethod for the calculation of high order energy/momentum aswell as amplitude dependencies of horizontal and vertical tune inthe storage ring of E989 using the differential algebra frameworkwithin COSY Infinity. First, the energy/momentum dependentreference orbit is calculated, which corresponds to the param-eter dependent fixed point of the map representing the detailedsimulations of the g-2 storage ring. Secondly, the fixed-pointmap is transformed into normal form coordinates to extract the

21

high-order tune dependencies. Analytical estimations of thosecalculations are presented to benchmark the simulation results.

——————————————–

MASSIVELY PARALLEL CHARGED PARTICLEOPTICS ENABLED BY MEMS FABRICATION

TECHNIQUES

Marco Wieland

Mapper Lithography

Invited TalkKeywords: maskless lithography, charged particle optics,MEMS, massively parallel

Mapper Lithography has developed a maskless lithography sys-tem, based on massively parallel electron-beam writing withhigh-speed optical data transport for switching the electronbeams. The system, containing 65,000 parallel electron beams,has a 1 wph throughput at 300 mm wafers and is capable of pat-terning any resolution and any different type of structure all theway down to 28 nm node patterns. The large number of beams isrealized by fabricating the electron optics using MEMS fabrica-tion techniques such as lithography and deep dry etching. In thispresentation we will discuss the various building blocks enabledby the MEMS fabrication techniques such as lens arrays, aperturearrays, deflector arrays, individually controllable deflectors andbeam blanker arrays. Also we will describe how these buildingblocks are combined to make a massively parallel electron-beamlithography tool.

——————————————–

TIME-OF-FLIGHT MASS SPECTROGRAPHS ANDTHE PRECISE DETERMINATION OF MASSES OF

IONS

H. Wollnik, M. Wada, P. Schury, M. Rosenbusch, Y. Ito,S. Kimura, H. Miyatake, S. Ishizawa

1 New Mexico State University, Las Cruces, NM 88001,USA; 2 RIKEN Nishina Center for Accelerator-Based

Science, Wako, Saitama 351-0198, Japan; 3 Wako NuclearScience Center (WNSC), Institute of Particle and Nuclear

Studies (IPNS); 4 High Energy Accelerator ResearchOrganization (KEK), Wako, Saitama 351-0198, Japan

5Institut of Physics, University of Tsukuba, Ibaraki305-8571, Japan

Invited TalkKeywords: Time-of-flight mass spectrographs, very precise massdeterminations, masses of short-lived nuclei

The masses of charged atoms and molecules were first in-vestigated by laterally dispersive sector field mass analyzers,which early on already achieved mass resolving powers m/∆m≈100 000 and more. Equally high mass resolving powerswere achieved by time-of-flight mass analyzers during the lastdecades. These measurements became possible when fast andprecise electronic circuitries became available. Such techniqueshave been developed and used extensively for the mass analysis

of short-lived nuclei, which mass values reveal insight in pro-cesses that describe the formation of elements in star explosions.Precise mass determinations of short-lived ions have been per-formed for energetic ions in large accelerator storage rings aswell as for low energy ions in time-of-flight mass spectrographswith long flight paths. Similarly precise mass measurements canalso performed for molecular ions that help to reveal the struc-ture of molecules. In case of very high mass resolving powersthe mass determination of molecular ions can be so high that themeasured ion mass directly reveals the molecule’s sum formula.

——————————————–

OPTIMIZE THE DESIGN OF CHICANES FOREEHG SCHEME

Kaishang Zhou

Shanghai Institute of Applied Physics

Keywords: CSR EEHG fully coherent

The echo-enabled harmonic generation (EEHG) scheme in freeelectron lasers is one of the most promise ways to generate thefully coherent soft x-rays. However, the increased energy spreadof the electron beam causing by the coherent synchrotron radi-ation (CSR) effect may smear the fine structure introduced bythe external seeded lasers. Here, we optimize the design of thechicanes to reduce the increased enegy spread causing by CSReffect.

——————————————–

22

ICAP’2018 ABSTRACTS

ZGOUBI: RECENT DEVELOPMENTS ANDFUTURE PLANS

Dan T. Abell (1), Rob Nagler (1), Francois Meot (2),Damian Rouson (3), and Izaak Beekman (4)

1. RadiaSoft, LLC, Boulder, CO; 2. Brookhaven NationalLab, Upton, NY; 3. Sourcery Institute, Oakland, CA; 4.

ParaTools, Inc., Eugene, OR

Classification: F-1, D-1, D-2

The particle tracking code Zgoubi [1,2] has been used for a broadarray of accelerator design studies, including FFAGs [3] andEICs [4,5]. Zgoubi is currently being used to evaluate proposeddesigns for both JLEIC [6,7] and eRHIC [8,9], and to preparefor commissioning the CBETA BNL-Cornell FFAG return loopERL [10,11,12]. Moreover, Zgoubi is now the subject of a PhaseII SBIR aimed at improving its speed, flexibility, and ease-of-use.In this paper, we describe our on-going work* on several fronts:(i) parallelizing Zgoubi using new features of Fortran 2018, in-cluding coarrays [13,14]; (ii) implementing a new particle updatealgorithm that requires significantly less memory and arithmetic;and (iii) developing symplectic tracking for field maps. In ad-dition, we describe plans for a web-based graphical interface toZgoubi. References 1. https://sourceforge.net/projects/zgoubi/2. F. Meot, FERMILAB-TM-2010, 1997 3. F. Lemuet et al.,NIM-A, 547:638, 2005 4. F. Meot et al., eRHIC/45, 2015 5.F. Lin et al., IPAC17, WEPIK114, 2017 6. J. Martinez-Marinet al., IPAC18, MOPMF004, 2018 7. A. M. Kondratenko etal., IPAC18, MOPML007, 2018 8. F. Meot et al., IPAC18,MOPMF013, 2018 9. V. H. Ranjbar et al., IPAC18, MOPMF016,2018 10. G. Hoffstaetter et al., IPAC18, TUYGBE2, 2018 11.F. Meot et al., NIM-A 896:60, 2018 12. F. Meot et al., FullField-Map Modeling of CBETA 4-Pass ERL, these proceed-ings 13. J. Reid, ISO/IEC JTC1/SC22/WG5 N2145, 2018 14.http://www.opencoarrays.org/ *This work was supported in partby the US Department of Energy, Office of Science, Office ofNuclear Physics under Award No. DE-SC0017181.

——————————————–

RECENT DEVELOPMENTS OF THE OPENSOURCE CODE OPAL

Andreas Adelmann for the OPAL developer team

PSI

Invited TalkClassification: F-1

After a general introduction of OPAL, I will introduce a set ofnew features available with version 2.0 released in July 2018. Allnew features will be presented together with examples of ongo-ing research projects. In the OPAL-t flavour, elements can nowbe placed in 3D, without restriction. Overlapping fringe fieldsare handled, and off-momentum beams as occurring in tolerance

studies can be tracked. Furthermore, survey plots of placed ele-ments are a valuable diagnostic when dealing with complex de-signs. A new element, a flexibly configurable collimator, willbe presented. In the OPAL-cyc flavour, a robust way of generat-ing matched distributions with linear space charge is introduced.A new method for describing fixed field accelerators (FFAs) ina very general way will be shown. A new element TRIMCOILcan be used to correct for field-errors in cyclotrons and FFAs.The OPAL language (a derivative of the MAD language) was ex-tended to allow the specification of multi objective optimisationproblems, which are then solved with a built in NGSA-II geneticalgorithm. A new feature SAMPLER allows you to setup and runrandom or sequential parameter studies and seamless utilisationof a vast number of computing cores. Finally, a set of Pythontools (pyOPALTools) was created for post processing. The man-ual is now available on the OPAL-wiki as well as in pdf format.

——————————————–

SNS BEAM TEST FACILITY FOR EXPERIMENTALBENCHMARKING OF HIGH INTENSITY BEAM

DYNAMICS COMPUTER SIMULATION

A. Aleksandrov, S. Cousineau, B. Cathey, A. Zhukov, Z.Zhang

ORNL, University of Tennessee

Classification: F-1, D-1

The SNS Beam Test Facility (BTF) is a 2.5 MeV hadron accel-erator equipped with state-of-the-art transverse and longitudinalbeam diagnostics. The BTF can produce pulsed high intensityH- beam with up to 50mA peak current. The expected availablebeam-on time of a few thousand hours per year provides an op-portunity for carrying out advanced high intensity beam dynam-ics experiments. The first ever direct measurement of 6D phasespace distribution of a beam in an accelerator has recently beencompleted. Preliminary analysis of the data shows a complexphase space structure that is not visible in measurements below5D, including correlations between degrees of freedom not cus-tomarily measured together. This result opens path forward tosolving the long-standing problem of initial condition in hadronlinac beam dynamics simulation. An extension of the BTF beamline consisting of a FODO line and high dynamic range emittancemonitor is being built to provide a test bench for simulation codesbenchmarking against measurements in well controlled environ-ment. This paper describes these efforts along with the longer-term plans.

——————————————–

23

QUANTUM STATISTICAL PROPERTIES OF FREEELECTRON LASER WITH A PLANAR WIGGLER

AND ION-CHANNEL GUIDING

Masoud Alimohamadi

Farhangian University

Classification: B-1

An analysis of the free-electron lasers (FELs) with a planar wig-gler and in the presence of ion-channel guiding, has been car-ried out using a Hamiltonian quantum field theory. The quan-tum Hamiltonian of single a particle has been derived in theBambini-Renieri (BR) frame [1-5]. The equations are valid ina reference frame, moving with a relativistic velocity with re-spect to the laboratory frame, chosen in such a way that the car-rier frequency of the pulse equals the pseudoradiation (wiggler)field frequency. In this reference frame, the equations assume asimple non-relativistic form. Time-dependent wave function andthree constants of motion are obtained. The Wei-Norman [2] Liealgebraic approach has been employed to solve exactly the spher-ical Raman-Nath equation (SRNE) [3-5]. A quantum approachhas been used to get photon gain, photon statistics and squeezingproperties of a FEL. The quantum statistical properties have alsobeen studied numerically. [1] H. Mehdian, M. Alimohamadi andA. Hasanbeigi, Journal of Plasma Physics 78 (5), 537-544(2012).[2] J. Wei, E. Norman, J. Math. Phys. A 4, 575 (1963). [3] M.Alimohamadi, H. Mehdian and A. Hasanbeigi, Journal of fusionenergy 31 (5), 463-466(2012). [4] A. Bambini and A. Renieri.Lett. Nuovo Cimento 21, 399 (1978). [5] F. Ciocci, G. Dattoli,A. Renieri and A. Torre, Physics Reports, 141(1), 1–50(1986).

——————————————–

OPTIMIZATION OF HEAVY-IONSYNCHROTRONS USING EVOLUTIONARYALGORITHMS AND MACHINE LEARNING

S. Appel, W. Geithner, S. Reimann, D. Vilsmeier, M.Sapinski, R. Singh

GSI

Invited TalkClassification: E-2

The application of machine learning and nature-inspired opti-mization methods, like for example genetic algorithms (GA)and particle swarm optimization (PSO) can be found in vari-ous scientific-technical areas. In accelerator physics these ap-proaches have not yet found a wide application. Still, in thelast years those approaches have been applied to a greater ex-tend. In this presentation, nature-inspired optimization as well asthe machine learning will be shortly introduced and their appli-cation to the accelerator facility at GSI/FAIR presented. For theheavy-ion synchrotron SIS18 at GSI the multi-objective GA/PSOoptimization resulted in a significant improvement of multi-turninjection performance and for the subsequent transmission forintense beams. A range of suitable injector brilliances for giveninitial loss could be defined. This information is crucial for thelayout of the injector upgrade for FAIR. The effect of transversespace charge force on MTI has been included in the optimiza-tion studies. An automated beam-setting optimization with ge-

netic algorithms at the CRYRING@ESR ion storage ring hasbeen performed. First results and the experience gained will bepresented. The application of machine learning for the recon-structing of space-charge distorted beam profiles from ionisationprofile monitors (IGMs) will be shown.

——————————————–

SELF-CONSISTENT SIMULATIONS OF SHORT-AND LONG-RANGE WAKEFIELD EFFECTS IN

STORAGE RINGS

Gabriele Bassi

Brookhaven National Laboratory

Invited TalkClassification: F-1

We discuss the parallel tracking code SPACE, which is capable tosimulate simultaneously the effect of short- and long range wake-fields on the dynamics of multi-bunch configurations in stor-age rings. As an example of such a simulation, we present astudy, performed at the NSLS-II storage ring, of the influenceof bunch lengthening and the microwave instability induced byshort-range wakefields, on the performance of a passive higherorder harmonic cavity for operation with multi-bunch configura-tions in hybrid modes.

——————————————–

SPIN DYNAMICS IN MODERN ELECTRONSTORAGE RINGS: COMPUTATIONAL ASPECTS

Oleksii Beznosov, Daniel Appelo, Desmond Barber,James Ellison, Klaus Heinemann

University of New Mexico, University of ColoradoBoulder, DESY, University of New Mexico, University of

New Mexico

Classification: A-2, F-1, F-2, D-2

In this talk we present some numerical results from our workon the spin polarization in high energy electron storage rings.The motivation of our work is to understand spin polarization invery high energy rings like the proposed Future Circular Collider(FCC-ee)[1] and Circular Electron Positron Collider (CEPC) [2].This talk is a supplement to K. Heinemann’s talk and gives fur-ther numerical details and results. As discussed in Heinemann’stalk our work is based on the initial value problem of the fullBloch equations (FBEs)[3] which in turn determines the polar-ization vector of the bunch. The FBEs take into account spin dif-fusion effects and spin-flip effects due to synchrotron radiation.The FBEs are a system of three uncoupled Fokker-Planck equa-tions plus coupling terms. Neglecting the spin flip terms in theFBEs one gets the reduced Bloch equations (RBEs) which posesthe main computational challenge. Our numerical approach hasthree parts. Firstly we approximate the FBEs analytically usingthe method of averaging, resulting in FBEs which allow us to uselarge time steps (without the averaging the time dependent coef-ficients of the FBEs would necessitate small time steps). Theminimum length of the time interval of interest is of the order ofthe orbital damping time. Secondly we discretize the averaged

24

FBEs in the phase space variables by applying the pseudospec-tral method, resulting in a system of linear first-order ODEs intime. The phase space variables come in d pairs of polar co-ordinates where d = 1, 2, 3 is the number of degrees of free-dom allowing for a d-dimensional Fourier expansion. The pseu-dospectral method is applied by using a Chebychev grid for eachradial variable and a uniform Fourier grid for each angle vari-able. Thirdly we discretize the ODE system by a time steppingscheme. The presence of parabolic terms in the FBEs necessi-tates implicit time stepping and thus solutions of linear systemsof equations. Dealing with 2d + 1 independent variables posesa computational challenge due to the extreme size of the ODEsystem if the Fourier modes are coupled extensively. However,thanks to having used averaged FBEs, the Fourier modes are un-coupled in the Fokker-Planck terms. Hence the parabolic termsare separated from the mode coupling terms. We take advan-tage of this separation by using an implicit/explicit time steppingscheme so that we end up with a large system of only locally cou-pled ODEs. Since the Fourier mode couplings are local, a par-allel implementation with only local communication is possible.Numerical experiments demonstrating efficiency and accuracy ofthe algorithm will be presented. References 1. FCC-ee webpagehttp://tlep.web.cern.ch 2. CEPC webpage http://cepc.ihep.ac.cn3. Ya.S.Derbenev, A.M. Kondratenko, “Relaxation and exqilib-rium state of electrons in storage rings”, Sov. Phys. Dokl. 19,p.438 (1975); K.A. Heinemann, O. Beznosov, J.A. Ellison, D.Appelo, D.P. Barber, “A Pseudospectral Method for Solving theBloch Equations of the Polarization Density in e- Storage Rings”,http://ipac2018.vrws.de/papers/thpak144.pdf

——————————————–

EXPLORING THE VALIDITY OF THE PARAXIALAPPROXIMATION FOR COHERENT

SYNCHROTRON RADIATION WAKE FIELDS

David Bizzozero, Erion Gjonaj, Herbert De Gersem

TU Darmstadt

Classification: C-2, F-1

Coherent synchrotron radiation (CSR) is an essential consider-ation in modern accelerators, yet is often computationally dif-ficult to accurately model. A common approach used in sim-ulating CSR effects uses the paraxial, or slowly-varying enve-lope approximation with a simple constant cross-section approx-imation of the geometry. While these approximations are of-ten valid for the simulation of many accelerator components, weaim to more closely analyze the errors introduced by such ap-proximations by comparing them with wake field solutions ob-tained by full-wave electromagnetic field simulations. The sim-ulations are performed with CSRDG (Coherent Synchrotron Ra-diation with Discontinuous Galerkin), our GPU-enabled MAT-LAB code. Presented in earlier work [Coherent Synchrotron Ra-diation and Wake Fields With Discontinuous Galerkin Time Do-main Methods, Proceedings of IPAC 2017, Copenhagen, Den-mark], CSRDG evolves Maxwell’s equations the time domainusing a curvilinear coordinate transformation and a Fourier se-ries decomposition in a transverse direction.

——————————————–

BEAM STABILITY ESTIMATES ANDSIMULATION STUDIES FOR THE FUTURE

CIRCULAR COLLIDER (FCC-HH)

Oliver Boine-Frankenheim, Daria Astapovych, UweNiedermayer, Sergey Arsenyev, Daniel Schulte

GSI, TU Darmstadt, CERN

Classification: A-2

Beam instabilities caused by impedances and electron clouds po-tentially limit the intensity and luminosity in the proposed Fu-ture Cicular Collider (FCC-hh). Scaling of the observed instabil-ity thresholds from the LHC to the FCC using simulation toolsis also one goal of the studies. Compared to the LHC the in-ner FCC beam screen radius is smaller and has two openings forthe synchrotron radiation. The complex beam pipe is the dom-inant contribution for beam instabilities. Using an impedancesolver in the frequency domain the pipe impedance is obtainedand instability growth rates are estimated. Besides the foreseenbroad-band damper system, the resistive wall induced transverseinstabilities should be stabilized by conventional and eventuallyalso advanced Landau damping concepts, which are studied us-ing particle tracking. Electron cloud buildup should be mitigatedin the FCC by either carbon or laser coating of the screen. Sim-ulations of electron cloud buildup including realistic secondaryemission yield (SEY) data and the detailed screen design are per-formed and and the resulting heat load and tune spreads are ana-lyzed.

——————————————–

BEAM DYNAMICS SIMULATIONS ANDCHALLENGES FOR THE FAIR SIS100

SYNCHROTRON

Oliver Boine-Frankenheim, Vera Chetvertkova,Vladimir Kornilov, Stefan Sorge, Yuan Yaoshuo

GSI, TU Darmstadt

Classification: D-1

The SIS100 synchrotron is the central accelerator of the upcom-ing FAIR project at GSI, Darmstadt, Germany. The major chal-lenges for the design studies and the later operation are relatedto high-intensity, low beam loss operation for a wide range ofion species and charge states, for different operational cycles andextraction schemes. We focus our simulation studies on the long(up to 1 s) accumulation plateau and on the final bunch compres-sion before extraction. During accumulation emittance growthand beam loss due to transverse space charge in combinationwith the magnet field errors has to be well controlled. We usedifferent simulation approaches with frozen and self-consistent“symplectic” space charge solvers to identify optimum workingpoint areas, including realistic field error models for the super-conducting, superferric SIS100 dipole and quadrupole magnets.

——————————————–

25

POLARIZED PROTON BEAMS FROMLASER-INDUCED PLASMAS

Markus Buescher, Anna Huetzen, Andreas Lehrach,Johannes Thomas

Forschungszentrum Juelich, Heinrich-Heine UniversityDuesseldorf

Classification: B-2, D-1, D-2, F-1

Laser-driven particle acceleration has undergone impressiveprogress in recent years. Nevertheless, one unexplored issueis how the particle spins are influenced by the huge magneticfields inherently present in the plasmas. In the framework of theJuSPARC (Juelich Short-Pulse Particle and Radiation Center) fa-cility and of the ATHENA consortium, the laser-driven genera-tion of polarized particle beams in combination with the devel-opment of advanced target technologies is being pursued. In or-der to predict the degree of beam polarization from a laser-drivenplasma accelerator, particle-in-cell simulations including spin ef-fects have been carried out for the first time. For this purpose, theThomas-BMT equation, describing the spin precession in elec-tromagnetic fields, has been implemented into the VLPL (VirtualLaser Plasma Lab) code. A crucial result of our simulations isthat a target containing pre-polarized hydrogen nuclei is neededfor producing highly polarized relativistic proton beams. For theexperimental realization, a polarized HCl gas-jet target is underconstruction the Forschungszentrum Juelich where the degree ofhydrogen polarization is measured with a Lamb-shift polarime-ter. The final experiments, aiming at the first observation of apolarized particle beam from laser-generated plasmas, will becarried out at the 10 PW laser system SULF at SIOM/Shanghai.

——————————————–

SPARSE GRIDS PARTICLE IN CELL SCHEMEFOR NOISE REDUCTION IN BEAM

SIMULATIONS

Antoine Cerfon

Courant Institute of Mathematical Sciences, New YorkUniversity

Invited TalkClassification: D-1, D-2, A-2

Kinetic simulations of intense charged particle beams are subjectto the curse of dimensionality: the run-time complexity of stan-dard solvers grows exponentially with the number of dimensionsof the underlying equations. This issue is particularly acute forcontinuum solvers, which need to discretize the six-dimensionalphase-space distribution function, and whose run times are con-sequently large even for a moderate number of grid points foreach dimension. Particle-in-Cell (PIC) schemes are a popularalternate approach to continuum methods, because they only dis-cretize the three-dimensional physical space and are thereforeless subject to the curse of dimensionality. Even if so, PICsolvers still have large run times, because of the statistical errorwhich is inherent to particle methods and which decays slowlywith the number of particles per cell. In this talk, we will presenta new scheme to address the curse of dimensionality and at thesame time reduce the numerical noise of PIC simulations. Our

PIC scheme is inspired by the sparse grids combination tech-nique, a method invented to reduce grid based error when solv-ing high dimensional partial differential equations [1]. The tech-nique, when applied to the PIC method, has two benefits: 1) italmost eliminates the dependence of the grid based error on di-mensionality, just like in a standard sparse grids application; 2) itlowers the statistical error significantly, because the sparse gridshave larger cells, and thus a larger number of particles per cellfor a given total number of particles. We will analyze the perfor-mance of our scheme for standard test problems in beam physics.We will demonstrate remarkable speed up for a certain class ofproblems, and less impressive performance for others. The latterwill allow us to identify the limitations of our scheme and exploreideas to address them. [1] Griebel M, Schneider M and ZengerC 1990 A combination technique for the solution of sparse gridproblems Iterative Methods in Linear Algebra ed R Bequwensand P de Groen (Amsterdam: Elsevier) pp 263-81

——————————————–

HIGH FIDELITY THREE-DIMENSIONALSIMULATIONS OF THERMIONIC ENERGY

CONVERTERS

Nathan Cook, Jon Edelen, Chris Hall, Mike Keilman,Paul Moeller, Rob Nagler

RadiaSoft LLC

Classification: E-1

Thermionic energy converters (TEC) are a class of thermoelec-tric devices, which promise improvements to the efficiency andcost of both small- and large-scale electricity generation. ATEC is comprised of a narrowly-separated thermionic emitterand an anode. Simple structures are often space-charge lim-ited as operating temperatures produce currents exceeding theChild-Langmuir limit. We present results from 3D simulationsof these devices using the particle-in-cell code Warp, developedat Lawrence Berkeley National Lab. We demonstrate improve-ments to the Warp code permitting high fidelity simulations ofcomplex device geometries. These improvements include mod-eling of non-conformal geometries using mesh refinement andcut-cells with a dielectric solver. We also consider self-consistenteffects to model Schottky emission near the space-charge limitfor arrays of shaped emitters. The efficiency of these devices iscomputed by modeling distinct loss channels, including kineticlosses, radiative losses, and dielectric charging. We demonstratemany of these features within an open-source, browser-based in-terface for running 3D electrostatic simulations with Warp, in-cluding design and analysis tools, as well as streamlined submis-sion to HPC centers.

——————————————–

26

UNCERTAINTY QUANTIFICATION FOR THEFUNDAMENTAL MODE SPECTRUM OF THE

EUROPEAN XFEL CAVITIES

Niklas Georg (1), Jacopo Corno (2), Herbert De Gersem(3), Shahnam Gorgi Zadeh (4), Ulrich Romer (1),

Sebastian Schops (2), Alexey Sulimov (5), Ursula vanRienen (4)

(1) Institute of Dynamics and Vibrations, TUBraunschweig; (2) Centre for Computational Engineering,TU Darmstadt; (3) TEMF, TU Darmstadt; (4) Theoretical

Electrical Engineering, Universitat Rostock; (5) DESY

Classification: E-2, C-2

The fundamental mode spectrum of superconducting cavities issensitive to small geometry deformations introduced by the man-ufacturing process. In this work we consider variations in theequatorial and iris radii of the 1.3 GHz TESLA cavities used atthe European XFEL. The cavities with slightly perturbed geom-etry are simulated using a finite element based eigenvalue solver.Employing uncertainty quantification methods such as sparse-grids, statistical information about the fundamental mode spec-trum can be efficiently calculated. Moreover, using global sen-sitivity analysis, in particular Sobol indices, the impact of theindividual geometry parameters on the quantities of interest, i.e.resonance frequencies, field-flatness and the cell-to-cell couplingcoefficient, can be computed. We will explain important aspectsof the uncertainty quantification methodology and give numeri-cal results for illustration.

——————————————–

HIGH-PRECISION LOSSY EIGENFIELDANALYSIS BASED ON THE FINITE ELEMENT

METHOD

Wolfgang Ackermann, Vinh Pham-Xuan, Herbert DeGersem

TEMF, TU Darmstadt, Germany

Classification: C-2

A proper eigenanalysis of resonating particle accelerator com-ponents is particularly advantageous to characterize structureswith high quality factors. While in former times eigenmodecalculations have been concentrating on the lossless cases only,meanwhile also lossy structures with finite-conductive materialsor with absorbing boundary conditions like PML or ports evenwith low quality factors are routinely available. In the losslesscase where no damping is present, all eigenvalues are locatedalong the real axis. If damping has to be modeled instead, thecorresponding eigenvalues are distributed within the first quad-rant of the complex plane that renders their determination muchmore expensive. One of the critical issues is that no resonanceshould be missed so that all desired eigenvalues in a given regionof the complex plane can be precisely determined. We imple-mented two different eigenvalue solvers based on a distributed-memory architecture. While the first one is a classical Jacobi-Davidson eigenvalue solver which has been adopted to be beused also within a complex-arithmetic environment, the second

one is based on the contour-integral method which enables to de-termine all eigenvalues within a given closed contour in the com-plex plane. Both solvers are attached to a FEM processor withsecond-order edge elements on curved tetrahedra and can be usedtogether in order to improve the computational efficiency. In thepresentation a selection of successful real-world applications ofthe implemented parallel eigenvalue solvers will be given.

——————————————–

SIXTRACK PROJECT: STATUS, RUNNINGENVIRONMENT AND NEW DEVELOPMENTS

R. De Maria, J. Andersson, V.K. Berglyd Olsen, L. Field,M. Giovannozzi, P.D. Hermes, N. Høimyr, S. Kostoglou,G. Iadarola, E.H. Maclean, E. Mcintosh, A. Mereghetti, J.

Molson, D. Pellegrini, T. Persson, M. Schwinzerl, K.Sjobak, I. Zacharov, S. Singh

CERN, Geneva, Switzerland; EPFL, Lausanne,Switzerland; IIT Madras, India

Classification: D-1

SixTrack is a single–particle tracking code for high–energy cir-cular accelerators routinely used at CERN for the Large HadronCollider (LHC), its luminosity upgrade (HL-LHC), the FutureCircular Collider (FCC), and the Super Proton Synchrotron(SPS) simulations. The code is based on a 6D symplectic track-ing engine, which is optimized for long–term tracking simula-tions and delivers fully reproducible results on many platforms.It also includes several scattering engines for beam–matter inter-actions studies, as well as facilities to run integrated simulationswith FLUKA and GEANT4. These features differentiate Six-Track from general–purpose, optics–design software like MAD-X. The code recently underwent a major restructuring to mergeadvanced features in a single branch such as multiple ion species,interface with external codes and high–performance input/output(XRootD, HDF5). In the process, the code moved from Fortran77 to Fortran 2018 standard, achieving also a better modulariza-tion. Physics models (beam–beam effects, rf–multipoles, currentcarrying wires, solenoid, electron–lenses) and methods (sym-plecticity check) have also been reviewed and refined to offermore accurate results. The SixDesk running environment allowsthe user to manage the large batches of simulations required foraccurate predictions of the dynamic aperture. SixDesk supportsCERN LSF and HTCondor batch systems, as well as the BOINCinfrastructure in the framework of the LHC@Home volunteer-ing computing project. SixTrackLib is a new library aimed atproviding a portable and flexible tracking engine for single– andmulti–particle problems using the models and formalism of Six-Track. The tracking routines are implemented in a parametrizedC code that is specialized to run vectorized in CPUs and GPUsusing SIMD intrinsics, OpenCL 1.2, and CUDA. This contribu-tion presents the status of the code and an outlook of future de-velopments of SixTrack, SixDesk and SixTrackLib.

——————————————–

27

BRAGG DIFFRACTION MODELING BETWEENX-RAY FREE-ELECTRON LASER AND CRYSTALS

Nanshun Huang, Kai Li, Haixiao Deng

Shanghai Institute of Applied Physics

Classification: B-1

In pursuit of fully coherent X-ray free-electron laser (FEL) [1],high reflective Bragg crystals have being and will be used as highselective spectral filter in the hard X-ray self-seeding FELs [2]and X-ray FEL oscillators (XFELO) [3], respectively. However,currently in the self-seeding FEL and XFELO simulations, thethree-dimensional effect of Bragg diffraction is not fully con-sidered. In this paper, we derive comprehensive solution forthe response function of crystal in Bragg diffraction. And athree-dimensional X-ray crystal Bragg diffraction code namedBRIGHT is introduced [4], which could collaborate closely withother FEL related code, e.g., GENESIS [5] and OPC [6]. Theperformance and feasibility are evaluated by two numerical ex-amples, i.e., self-seeding experiment for LCLS [7] and XFELOoptions for Shanghai high repetition rate XFEL and extreme lightfacility (SHINE) [8]. The results indicate BRIGHT provides anew and useful tool for three-dimensional FEL simulation. [1]R. Bonifacio, C. Pellegrini, and L. M. Narducci, Collective insta-bilities and high-gain regime in a free electron laser, Opt. Com-mun. 50, 373 (1984). [2] J. Amann, et al., “Demonstration ofself-seeding in a hard-x-ray free-electron laser,” Nat. Photonics6, 693–698 (2012). [3] K. J. Kim, Y. Shvydko, and S. Reiche,“A proposal for an x-ray free-electron laser oscillator with anenergy-recovery linac,” Phys. Rev. Lett. 100, 244802 (2008).[4] N. Huang, K. Li, H. Deng, BRIGHT: the three-dimensionalX-ray crystal Bragg diffraction code (In preparation) [5] S. Re-iche, “Genesis 1.3: A fully 3d time-dependent FEL simulationcode,” Nucl. Instrum. Methods Phys. Res., Sect. A 429, 243–248 (1999). [6] P. J. M. van der Slot, H. P. Freund, W. H. Miner,Jr., S. V. Benson, M. Shinn, and K.-J. Boller, “Time-dependent,three-dimensional simulation of free-electron-laser oscillators,”Phys. Rev. Lett. 102, 244802 (2009). [7] P. Emma, R. Akre, J.Arthur, R. Bionta, C. Bostedt, J. Bozek, A. Brachmann, P. Bucks-baum, R. Coffee, F.-J. Decker et al., “First lasing and operationof an angstrom-wavelength free-electron laser,” Nat. Photonics4, 641–647 (2010). [8] K. Li, H. Deng, Systematical designand three-dimensional simulation of X-ray FEL oscillator for theShanghai coherent light facility, Nucl. Instr. and Meth. A, 895(2018) 40-47.

——————————————–

UPGRADE OF MAD-X FOR HL-LHC PROJECTAND FCC STUDIES

Laurent Deniau, Helmut Burkhardt, MassimoGiovannozzi, John M. Jowett, Andrea Latina, Tobias

Persson, Frank Schmidt, and Piotr Krzysztof Skowronski

CERN

Classification: A-2, D-1, F-1

The design efforts for the High Luminosity upgrade of the LargeHadron Collider (HL-LHC) and for the FCC-ee project requiredsignificant extensions of the MAD-X code widely used for de-signing and simulating particle accelerators. The modelling ofsynchrotron radiation effects has recently been reviewed, im-proved and tested on the lattices of ESRF, LEP and CLIC Fi-nal Focus System. The results were cross checked with thecodes AT, PLACET, Geant4, and MAD8. The implementationof space charge has been drastically restructured in a modulardesign. The linear coupling calculation has been completely re-viewed and improved, from the theory to the implementation inMAD-X code to ensure its correctness in the presence of strongcoupling as in the HL-LHC studies. The slicing module hasbeen generalised to allow for thick slices of bending magnets,quadrupoles and solenoids. The SBEND element has been ex-tended to support difference between bending angle and inte-grated dipole strength. Patches have been added to the list of sup-ported elements. MAD-X PTC has also been extended to trackresonance driving terms along layouts, and to support AC dipolesto simulate beams during optics measurements.

——————————————–

SURROGATE MODELS FOR BEAM DYNAMICS INCHARGED PARTICLE ACCELERATORS

Auralee Edelen, Andreas Adelmann, Nicole Neveu,Matthias Frey, Dinesh Acharya

SLAC, PSI, ANL, PSI, PSI

Classification: E-2, F-2, D-1

High-fidelity, PIC-based beam dynamics simulations are timeand resource intensive. Consider a high dimensional searchspace, that is far too large to probe with such a high resolu-tion simulation model. We demonstrate that a coarse samplingof the search space can produce surrogate models, which are ac-curate and fast to evaluate. In constructing the surrogate mod-els, we use artificial neural networks [1] and multivariate poly-nomial chaos expansion [2]. The performance of both meth-ods are demonstrated in a comparison with high-fidelity simu-lations, using OPAL, of the Argonne Wakefield Accelerator [3].We claim that such surrogate models are good candidates for ac-curate on-line modeling of large, complex accelerator systems.[1] A. L. Edelen et al., arXiv:1610.06151[physics.acc-ph] [2] A.Adelmann, arXiv:1509.08130v6[physics.acc-ph] [3] N. Neveu etal., 2017 J. Phys.: Conf. Ser. 874 012062

——————————————–

28

NORMAL FORM APPROACH TO ANDNONLINEAR OPTICS ANALYSIS OF THE IOTA

RING

Bela Erdelyi

Northern Illinois University

Invited TalkClassification: D-1, A-2

The IOTA ring is the realization as an accelerator system ofa nonlinear, completely integrable Hamiltonian. Normal formmethods allow analysis of one-turn maps of rings, exposingglobal information about the dynamics, including amplitude de-pendent tune shifts and resonance strengths. Since mapping thephase space of particle dynamics in IOTA is important to gaininsight and offer practical ways to optimize for intensity frontierbeam physics, this talk will summarize our group’s results, theadvantages, difficulties, and limitations of normal form analysisof the IOTA nonlinear optics.

——————————————–

ELECTRON BEAM LONGITUDINAL PHASESPACE RESTORATION FROM THE IMAGE

AFTER BEAM PASS DEFLECTOR CAVITY ANDSPECTROMETER ARM

Mikhail Fedurin

Brookhaven National Laboratory, Accelerator Test Facility

Classification: D-2

Recently commissioned X-band deflector cavity at BrookhavenNational Laboratory Accelerator Test Facility (BNL ATF) is usedfor electron bunch longitudinal profile measurements in both - atzero-degree beamline and at spectrometer arm directions to mea-sure the e-beam longitudinal phase space profile. The deflec-tor cavity induces energy distortions on the off-axis particles andcorrupt real picture of the beam energy profile at spectrometerscreen. A special computational phase space restoration tech-nique which is under development at BNL ATF to reveal undis-torted e-beam parameters will be discussed.

——————————————–

COMPUTATION OF EIGENMODES IN LONG ANDCOMPLEX ACCELERATING STRUCTURES BY

MEANS OF CONCATENATION STRATEGIES

Thomas Flisgen

Helmholtz-Zentrum Berlin

Invited TalkClassification: C-2

The computation of eigenmodes in chains of superconductingcavities with asymmetric couplers is a demanding problem. Thisproblem typically requires the use of high-performance comput-ers in combination with dedicated software packages. Alterna-tively, the eigenmodes of chains of superconducting cavities canbe determined by the so-called State-Space Concatenation (SSC)

approach that has been developed at the University of Rostock.SSC is based on the decomposition of the full chain into individ-ual segments. Subsequently, the RF properties of every segmentare described by reduced order models. These reduced ordermodels are concatenated to a reduced order model of the entirechain by means of algebraic side constraints arising from con-tinuity conditions of the fields across the decomposition planes.The constructed reduced order model describes the RF propertiesof the complete structure so that the field distributions, the cou-pling impedances and the external quality factors of the eigen-modes of the full cavity chain are available. In contrast to directmethods, SSC allows for the computation of the eigenmodes ofcavity chains using desktop computers. The current contribu-tion discusses theoretical aspects of the scheme. Moreover, a setof application examples is presented such as eigenmode com-putations for cavity chains of the European XFEL, BERlin Pro,and BESSY VSR. Furthermore, an outlook is presented to in-clude surface losses into the formulation by using perturbationapproaches.

——————————————–

TRIMCOIL OPTIMISATION USINGMULTI-OBJECTIVE OPTIMISATION

TECHNIQUES AND HPC

Matthias Frey, Jochem Snuverink, Andreas Adelmann

Paul Scherrer Institut (PSI)

Classification: D-1, E-2, A-2, F-1, F-2

Uncertainties in the bunch injection (i.e. energy, radius, radialmomentum and angle) as well as magnet inaccuracies harm theisochronicity of the PSI 590 MeV Ring Cyclotron. An additionalmagnetic field provided by trim coils is an effective solutionto restore this condition. Therefore, an accurate description oftrim coils is essential to match the turn pattern of the machinein simulations. However, due to the high-dimensional searchspace consisting of 21 design variables and more than 180 ob-jectives the turns cannot be matched in a straightforward mannerand without sufficient HPC resources. In this talk we present arealistic trim coil model for the PSI 590 MeV Ring Cyclotronimplemented in OPAL that was used together with its built-inmulti-objective optimisation algorithm to find the 4 injection pa-rameters and the magnetic field strengths of 17 trim coils. Theoptimisations were performed on Piz Daint (currently 3rd fastestsupercomputer world-wide) with more than 1000 cores per job.

——————————————–

29

COMPUTER ARCHITECTURE INDEPENDENTADAPTIVE GEOMETRIC MULTIGRID SOLVER

FOR AMR-PIC

Matthias Frey, Andreas Adelmann

Paul Scherrer Institut (PSI)

Classification: F-2, F-1, D-1, D-2

The accurate and efficient simulation of neighboring buncheffects in high intensity cyclotrons requires to solve large-scale N-body problems of O(109...1010) particles coupled withMaxwell’s equations. In order to capture the effects of halo cre-ation and evolution of such simulations with standard particle-in-cell models an extremely fine mesh with O(108...109) gridpoints is necessary to meet the condition of high resolution. Thisrequirement represents a waste of memory in regions of void,therefore, the usage of block-structured adaptive mesh refine-ment algorithms is more suitable. The N-body problem is thensolved on a hierarchy of levels and grids using geometric multi-grid algorithms. We show benchmarks of a new implementationof an adaptive geometric multigrid algorithm using 2nd gener-ation Trilinos packages that ran on Piz Daint with O(104...105)cores.

——————————————–

RECENT DEVELOPMENTS IN WAKE FIELD ANDBEAM DYNAMICS COMPUTATION

Erion Gjonaj, David A. Bizzozero, Steffen Schmid,Herbert De Gersem

Darmstadt University of Technology, ComputationalElectromagnetics Laboratory

Plenary TalkClassification: C-2, F-1, D-1, B-1

Wake potentials and beam coupling impedances can be calcu-lated analytically only for simple structures and special limitingcases. For the calculation of wake fields in “real-world” 3D ac-celerator structures, one has to rely on numerical field computa-tions. Among the most successful numerical techniques for wakefield calculations in the time domain are dispersion-free methodsin the moving window. These techniques are particularly use-ful for short-range wake field calculations. Recently, this classof methods has been extended to include Surface ImpedanceBoundary Conditions (SIBC) based on the Auxiliary Differen-tial Equation (ADE) technique. These boundary conditions allowthe computation of resistive wall wake fields for 3D structureswith arbitrary frequency dependent conductivity. An importantapplication of this method is the calculation resistive wall wakefields in novel accelerator chambers with NEG and amorphouscarbon coatings. Other developments to be discussed includethe calculation of CSR-wakes in bunch compressors and undu-lator structures for x-ray sources. This task is computationallyvery difficult because of the curved bunch trajectory that leadsto extremely high frequency and long-range wake fields. Timedomain as well as frequency domain methods based on high or-der DG and FE discretization techniques for the electromagneticfields computation in such structures will be presented.

——————————————–

EXPERIENCE WITH CBETA ONLINE MODELINGTOOLS

C. Gulliford, D. Sagan, A. Bartnik, J. Dobbins, J.S. Berg

Cornell University

Classification: A-2, B-1, D-1, E-1, D-2, E-2, F-1

The CBETA machine is a four pass Energy Recovery Linac(ERL) with an Fixed-Field Alternating Gradient (FFAG) arccurrently being developed as a joint project between the Cor-nell Laboratory for Accelerator-Based Sciences And Education(CLASSE) and Brookhaven National Lab. For online modelingof CBETA, a customized version of the Tao program, which isbased upon the Bmad toolkit, is used along with the GPT pro-gram for low energy space charge calculations. The customizedversion of Tao, called CBETA-V, is interfaced to python for com-munication with the EPICS control system. This paper describesthe online modeling system and initial experience during ma-chine running.

——————————————–

SIMULATIONS OF BEAM CHOPPING FORPOTENTIAL UPGRADE OF THE SNS LEBT

CHOPPER

B.X. Han, R.F. Welton, V. Peplov, R. Saethre, S.N.Murray Jr., T.R. Pennisi, C.M. Stinson, and M.P. Stockli

Spallation Neutron Source, Oak Ridge NationalLaboratory, Oak Ridge, TN 37831, USA

Classification: D-1, A-2, E-2

The Spallation Neutron Source (SNS) accelerator system in-cludes a 65 keV H- injector, a 2.5 MeV RFQ, a 1 GeV linacchain (DTL-CCL-SCL), and an accumulator ring. The H- in-jector consists of a RF-driven, Cs-enhanced H- ion source anda two-lens electrostatic low energy beam transport (LEBT) thatfeed the RFQ accelerator with 1 ms H- beams pulsed at 60 Hz.To facilitate the multi-turn beam stacking in the ring and to createa gap for clean beam extraction from the ring, the H- beams arechopped in the LEBT section in front of the RFQ at the ring revo-lution frequency (∼1 MHz). The second lens of the LEBT is az-imuthally split into four segments to allow applications of varioustransverse electric fields for beam steering, chopping or blanking.Currently, the four segments are pulsed independently by fourbipolar high voltage pulse generators and the four pulse gener-ators are powered simultaneously at a time to chop the beamwith ∼1 MHz repetition rate and toward the four different di-agonal directions sequentially. In addition to a plan for upgrad-ing the nearly obsolete high voltage pulse generators, differenttiming configurations for beam chopping are being proposed toimprove the pulse generator performance by reducing switchingfrequency and power dissipation in the high voltage pulse gen-erators. New chopping configurations where only two segmentsor even only one segment is used at a time are proposed. Thispaper presents simulations of the beam behavior under these new

30

chopping configurations to evaluate the beam chopping perfor-mance including the required high voltage amplitudes to deflectthe beam out of the RFQ acceptance in phase-space, and the dis-tributions of the deflected beams on the LEBT chopper target.

——————————————–

DESIGN STUDY OF A FAST KICKER MAGNETAPPLIED TO THE BEAMLINE OF A PROTON

THERAPY FACILITY

Wenjie Han, Bin Qin, Jun Yang, Kaifeng Liu, ZhikaiLiang, Xu Liu

Huazhong University Of Science And Technology

Classification: C-1

A proton therapy facility based on isochronous superconduct-ing cyclotron is under development in HUST (Huazhong Uni-versity of Science and Technology). A fast kicker magnet willbe installed at the upstream of the degrader to perform the beamswitch function by kicking the proton beam to the downstreambeam stop. The rising and falling time of the kicker is about100us and the maximum repetition rate is 500Hz. This paperintroduces simulation and optimization of the eddy current anddynamic magnetic field of the fast kicker, by using FEM codeOPERA-3D. For kicker materials, laminated steel and ferrite arecompared and the MnZn ferrite was chose. Design considera-tions including the eddy current effect, field hysteresis and me-chanical structure of the kicker will also be introduced with mul-tiphysics analysis.

——————————————–

SPIN DYNAMICS IN MODERN ELECTRONSTORAGE RINGS: COMPUTATIONAL AND

THEORETICAL ASPECTS

Klaus Heinemann, Daniel Appelo, Desmond P. Barber,Oleksii Beznosov, James A. Ellison

University of New Mexico, University of ColoradoBoulder, Deutsches Elektronen Synchrotron, University of

New Mexico, University of New Mexico

Plenary TalkClassification: D-2, D-1, F-1, F-2, A-2

In this talk we present some numerical and analytical resultsfrom our work on the spin polarization in high energy electronstorage rings. Our work is based on the initial value problemof what we call the full Bloch equations (FBEs). The solutionof the FBEs is the polarization density which is proportional tothe spin angular momentum density per particle in phase spaceand which determines the polarization vector of the bunch. TheFBEs take into account spin diffusion effects and spin-flip ef-fects due to synchrotron radiation including the Sokolov-Ternoveffect and its Baier-Katkov generalization. The FBEs were intro-duced by Derbenev and Kondratenko in 1975 as a generalizationof the Baier-Katkov-Strakhovenko equations from a single orbitto the whole phase space. The FBEs are a system of three un-coupled Fokker-Planck equations plus two coupling terms, i.e.,the T-BMT term and the Baier-Katkov term. Neglecting the spin

flip terms in the FBEs one gets what we call the reduced Blochequations (RBEs). The RBEs are sufficient for computing the de-polarization time. The conventional approach of estimating andoptimizing the polarization is not based on the FBEs but on theso-called Derbenev-Kondratenko formulas. However, we believethat the FBEs offer a more complete starting point for very highenergy rings like the FCC-ee and CEPC. The issues for very highenergy are: (i) Can one get polarization, (ii) are the Derbenev-Kondratenko formulas satisfactory at very high energy? If not,what are the theoretical limits of the polarization? Item (ii) willbe addressed both numerically and analytically. Our numericalapproach has three parts. Firstly we approximate the FBEs ana-lytically using the method of averaging, resulting in FBEs whichallow us to use large time steps (without the averaging the timedependent coefficients of the FBEs would necessitate small timesteps). The minimum length of the time interval of interest is ofthe order of the orbital damping time. Secondly we discretizethe averaged FBEs in the phase space variables by applying thepseudospectral method, resulting in a system of linear first-orderODEs in time. The phase space variables come in d pairs of po-lar coordinates where d=1,2,3 is the number of degrees of free-dom allowing for a d-dimensional Fourier expansion. The pseu-dospectral method is applied by using a Chebychev grid for eachradial variable and a uniform Fourier grid for each angle vari-able. Thirdly we discretize the ODE system by a time steppingscheme. The presence of parabolic terms in the FBEs necessi-tates implicit time stepping and thus solutions of linear systemsof equations. Dealing with 2d+1 independent variables posesa computational challenge due to the extreme size of the ODEsystem if the Fourier modes are coupled extensively. However,thanks to having used averaged FBEs, the Fourier modes are un-coupled in the Fokker-Planck terms. Hence the parabolic termsare separated from the mode coupling terms. We take advan-tage of this separation by using an implicit/explicit time steppingscheme so that we end up with a large system of only locallycoupled ODEs. Since the Fourier mode couplings are local, aparallel implementation with only local communication is pos-sible. Some numerical results will be shown. Details and moreresults will be presented in the talk by O.Beznosov.

——————————————–

OPTIMIZATION OF HADRON THERAPYBEAMLINES USING A NOVEL FAST TRACKING

CODE FOR BEAM TRANSPORT ANDBEAM-MATTER INTERACTIONS

Cedric Hernalsteens (1,2), Kevin Andre (1), VincentCollignon (1), Quentin Flandroy (1), Baptiste Herrengods(1), Raphael Jungers (2), Robin Tesse (3), Zheming Wang

(2)

(1) Ion Beam Applications (IBA) (2) UniversiteCatholique de Louvain (3) Universite Libre de Bruxelles

Classification: E-2, F-1, A-2, E-1

The optimization of proton therapy beamlines challenges the tra-ditional approach used in beam optics due to the very strict con-straints on beam quality, especially for Pencil Beam Scanning,despite the large losses induced by the emittance increase com-ing from the energy degrader. In order to explore the perfor-

31

mances of proton therapy beamlines, we proceed using a newfast beam tracking Python library coupled with a genetic algo-rithm. Global optimization algorithms such as the genetic algo-rithm or basin hopping schemes require numerous evaluations ofthe model and their practical implementations are limited by thecomputation time at each iteration. To overcome this limitation,while at the same time allowing an open-box user experience,a Python library has been developed, including transport mod-els for the typical hadron therapy beamlines elements, as well asmodels for the computation of multiple Coulomb scattering. TheMulti-Objective Genetic Algorithm (MOGA) allows to explorethe parameter space in a global sense. This multi-objective al-gorithm enables the simultaneous optimization of complex con-straints specific to proton therapy beamlines. Results for the IBAProteus One system are presented and discussed in detail.

——————————————–

NOVEL, FAST, OPEN-SOURCE CODE FORSYNCHROTRON RADIATION COMPUTATION ON

ARBITRARY 3D GEOMETRIES

Dean Andrew Hidas

Brookhaven National Laboratory

Classification: B-1, C-2

Open Source Code for Advanced Radiation Simulation (OS-CARS) is an open-source project (https://oscars.bnl.gov) devel-oped at Brookhaven National Laboratory for the computation ofsynchrotron radiation from arbitrary charged particle beams inarbitrary and time-dependent magnetic and electric fields on ar-bitrary geometries in 3D. Computational speed is significantlyincreased with the use of built-in multi-GPU and multi-threadedtechniques which are suitable for both small scale and large scalecomputing infrastructures. OSCARS is capable of computingspectra, flux, and power densities on simple surfaces as well as onobjects imported from common CAD software. It is additionallyapplicable in the regime of high-field acceleration. The method-ology behind OSCARS calculations will be discussed along withpractical examples and applications to modern accelerators andlight sources.

——————————————–

DESIGN AND SIMULATION OF HIGHMOMENTUM ACCEPTANCE GANTRIES FOR ION

BEAM THERAPY

Anthony Huggins, Weishi Wan, Lucas Brouwer

Varian Medical Systems, University of Duesseldorf,LBNL, PSI

Invited TalkClassification: A-2, D-1

One challenge of proton beam therapy is the shear size of itsequipment. A proton gantry that rotates a beamline about a pa-tient is typically about 10 meters in diameter, heavy and expen-sive. One approach to reduce size and cost of gantries is theirminiaturization by the application of superconducting (SC) mag-nets in the beamline. SC magnets, however, have difficulties

to quickly adapt their field when the beam energy is changed.Achromatic beamline designs with high momentum acceptancebased on superconducting magnets can lead to compact gantriesthat still allow rapid beam application which is an important clin-ical requirement. In a collaborative effort LBNL, Varian Medi-cal Systems and PSI have developed the Alternating GradientCanted-Cosine-Theta (AG-CCT), a curved version of the CCTdesign that includes alternating quadrupole and sextupole com-ponents to build an achromat. The AG-CCT reaches a momen-tum acceptance of approx. 20 % dp/p while preserving beamprofiles within clinical specification. Another design, conceivedby LBNL and Varian, achieves momentum acceptance over theentire clinical beam energy range (70-225 MeV), called the fixed-field achromat. The beam optics principles of the two achromatsand an optimized associated gantry beamline design is the mainfocus of the presented work, as well as putting these in contextof clinical requirements and economic constraints.

——————————————–

FEL SIMULATION USING THE LIE METHOD

Kilean Hwang, Ji Qiang

Lawrence Berkeley National Lab

Invited TalkClassification: B-1

Advances in numerical methods for free-electron-laser (FEL)simulation under wiggler period averaging (WPA) are presented.First, WPA is generalized using perturbation Lie map method.The conventional WPA is identified as the leading order contri-bution. Next, shot-noise model under WPA is improved alongwith a particle migration scheme across the numerical mesh. Theartificial shot noise arising from particle migration across numer-ical mesh is suppressed. The improved model also allows usingarbitrary mesh size, slippage resolution, and integration step size.These advances will improve modeling of longitudinal beam pro-file evolution for fast FEL simulation.

——————————————–

ANALYSIS OF THE BEAM LOSS MECHANISMDURING THE ENERGY RAMP-UP AT THE

SAGA-LS

Yoshitaka Iwasaki

SAGA Light Source

Classification: B-1

The accelerator of the SAGA Light Source consists of 255 MeVinjector linac and 1.4 GeV storage ring. The accumulated elec-tron beam current of the storage ring is about 300 mA. The en-ergy of the electrons are raised up to 1.4 GeV in 4 minutes inthe storage ring. At the moment of the beam acceleration (thebeam energy is lower than 300 MeV), the electron beam is lostlike the step function. The lost beam current is normally about5 mA to 30 mA. The beam loss at the energy ramp-up is notobserved, when the beam current is lower than 200 mA. To un-derstand the beam loss mechanism, which depend on the beamcurrent, we developed high-speed logging system of 100 kHz for

32

monitoring the beam current and the magnets power supplies us-ing National Instruments PXI. We investigated the relationshipbetween the beam loss and the betatron tune shifts. The tuneshifts during the beam acceleration were analyzed from the mea-sured data of the output current of the magnets power suppliesby using beam tracking code of TRACY2. By adopting the newhigh-speed logging system, the time structure of the beam lossprocess was clearly observed. We will present the high-speedlogging system developed for monitoring the beam current andthe power supplies at this meeting. The results of the investiga-tion to find the relationship of the beam loss and the tune shiftswill be also shown.

——————————————–

SINGLE OBJECTIVE GENETIC OPTIMIZATIONOF AN 85% EFFICIENT KLYSTRON

Aaron Jensen, John Petillo, Lawrence Ives, MichaelRead, Jeff Neilson

Leidos, Calabazas Creek Research, SLAC NationalAccelerator Laboratory

Classification: E-2, C-2, E-1

Overall efficiency is a critical priority for the next generation ofparticle accelerators as they push to higher and higher energies.In a large machine, even a small increase in efficiency of anysubsystem or component can lead to a significant operational costsavings. The Core Oscillation Method (COM) and Bunch-Align-Compress (BAC) method have recently emerged as a means togreatly increase the efficiency of the klystron RF source for par-ticle accelerators. The COM and BAC methods both work byuniquely tuning klystron cavity frequencies such that more parti-cles from the anti-bunch are swept into the bunch before power isextracted from the beam. The single objective genetic algorithmfrom Sandia National Laboratory’s Dakota optimization libraryis used to optimize both COM and BAC based klystron designsto achieve 85% efficiency. The COM and BAC methods are dis-cussed. Use of the Dakota optimization algorithm library fromSandia National Laboratory is discussed. Scalability of the op-timization approach to High Performance Computing (HPC) isdiscussed. The optimization approach and optimization resultsare presented.

——————————————–

ADVANCED DESIGN AND SIMULATION OFFIXED-FIELD ACCELERATORS

Carol Johnstone, Martin Berz, Kyoko Makino, PavelSnopok

Particle Accelerator Corp., Michigan State Univ.,Michigan State Univ., Illinois Institute of Technology

Classification: A-2, F-1

The development of new types of accelerators that allow widechoices of parameters, promote complicated fields, and oftenneed to efficiently handle very large emittance beams requiresthe availability of new simulation environments to design and

accurately predict operation. This is particularly true of Fixed-field accelerators, FFAs, which apply arbitrary-order fields - bothalternating gradient, strong focusing - but also weak-focusingcyclotrons. This is especially applicable at medium-to-high en-ergy combined with high intensity (mA currents). Synchrotronand cyclotron codes are generally inadequate to simulate accu-rately the performance of these strong-focusing fixed-field ac-celerators, particularly the new breed of non-scaling machineswhich have difficult, high-order fringe-field and edge-angle ef-fects. One well-supported code, COSY INFINITY (COSY) isparticularly suitable for accurate, high-order descriptions of ac-celerators. New tools have been developed in COSY INFINITYto address and accurately represent complex fixed-field machinesin both a sector and spiral sector footprint. A description, appli-cation, and comparison of these tools with fields from magnetlattice design is presented.

——————————————–

S-BASED MULTI-PARTICLE SPECTRALSIMULATION OF AN ELECTRON GUN

Paul M. Jung, Thomas Planche, Rick Baartman

TRIUMF

Classification: A-2, D-1, D-2

We derive a Hamiltonian description of a continuous particle dis-tribution and its electrostatic potential from the Low Lagrangian.The self consistent space charge potential is discretized accord-ing to the spectral Galerkin approximation. The particle distri-bution is discretized using macro-particles. We choose a set ofinitial and boundary conditions to model the TRIUMF 300keVthermionic DC electron gun. The field modes and macro-particlecoordinates are integrated self-consistently using map methods.The results are compared to results obtained from ASTRA simu-lations and experimental data.

——————————————–

HOM-MITIGATION FOR FUTURE SPS 33-CELL200 MHZ ACCELERATING STRUCTURES

P. Kramer, C. Vollinger

CERN

Classification: C-2, A-2, E-1

The CERN SPS 200 MHz travelling wave (TW) acceleratingstructures pose an intensity limitation for the planned high lumi-nosity (HL-) LHC upgrade. Higher-order modes (HOMs) around630 MHz have been identified as one of the main sources oflongitudinal multi-bunch instabilities. Improved mitigation ofthese HOMs with respect to today’s HOM-damping scheme istherefore an essential part of the LHC injectors upgrade (LIU)project. The principles of HOM-couplers in cavities and thepresent damping scheme are reviewed, before illustrating the nu-merous requirements an improved damping scheme for the future33-cell structures must fulfil. These are, amongst others, the mit-igation of HOMs situated in the lower part of the structure wherethere are no access ports for extraction, a sufficient overall damp-ing performance and an acceptable influence on the fundamental

33

accelerating passband (FPB). Different approaches tackling thesechallenges are investigated by 3D electromagnetic (EM) simula-tions and their performance, advantages and pitfalls are evalu-ated. The most promising solution involves slight changes of theaccelerating structure, e.g. by the insertion of metallic perturbersin the bottom part of the cavity. Their behaviour is confirmed bylab measurements and its compatibility with the travelling-waveFPB is further examined.

——————————————–

CONSTRAINED MULTI-OBJECTIVE SHAPEOPTIMIZATION OF SUPERCONDUCTING RF

CAVITIES TO COUNTERACT DANGEROUSHIGHER ORDER MODES

Marija Kranjcevic, Shahnam Gorgi Zadeh, AndreasAdelmann, Peter Arbenz, Ursula van Rienen

ETH Zurich, University of Rostock, Paul Scherrer Institut(PSI), ETH Zurich, University of Rostock

Classification: C-2, F-2, E-2

High current storage rings, such as the Z operating mode of theFCC-ee, require accelerating cavities that are optimized with re-spect to both the fundamental mode and the dangerous higherorder modes (HOMs). In such cavities, monopole and dipolemodes are the major sources of beam instability and have to besufficiently damped. In addition to the damping method, opti-mizing the shape of the superconducting radio frequency (RF)cavity can help lower the effect of dangerous HOMs. In order tooptimize the shape of the RF cavity we solve a constrained multi-objective optimization problem using a massively parallel imple-mentation of an evolutionary algorithm. Focusing on axisym-metric RF cavities, we parameterize and mesh their cross sec-tion, and then use a fast 2D Maxwell eigensolver to solve time-harmonic Maxwell’s equations. For each cavity, after the Fourierexpansion in the azimuthal direction, we need to solve the eigen-problems corresponding to the few lowest Fourier modes. Weinvestigate various approaches to parallelize this and implementa repair method to deal with the constraint on the frequency ofthe fundamental mode. Finally, we show the computed Paretofront approximation and individuals with good objective func-tion values, i.e. the RF cavity shapes with desired properties.

——————————————–

START-TO-END SIMULATIONS OF THZ SASE FELPROOF-OF-PRINCIPLE EXPERIMENT AT PITZ

Mikhail Krasilnikov, Prach Boonpornprasert

DESY, Zeuthen, Germany

Classification: B-1, D-1

The Photo Injector Test facility at DESY in Zeuthen (PITZ) de-velops high brightness electron sources for modern linac-basedFree Electron Lasers (FELs). The PITZ accelerator has beenproposed as a prototype for a tunable, high power THz sourcefor pump and probe experiments at the European XFEL. A Self-Amplified Spontaneous Emission (SASE) FEL is considered togenerate the THz pulses. High radiation power can be achieved

by utilizing high charge (4 nC) shaped electron bunches fromthe PITZ photo injector. THz pulse energy of up to several mJis expected from preliminary simulations for 100 um radiationwavelength. For the proof-of-principle experiments a re-usage ofLCLS-I undulators at the end of the PITZ beamline is under stud-ies. One of the challenges for this setup is transport and matchingof the space charge dominated electron beam through the narrowvacuum chamber. Start-to-end simulations for the entire experi-mental setup - from the photocathode to the SASE THz genera-tion in the undulator section - have been performed by combina-tion of several codes: ASTRA, SC and GENESIS-1.3. The spacecharge effect and its impact onto the output THz radiation havebeen studied. The results of these simulations will be presentedand discussed.

——————————————–

REVIEW OF SPECTRAL MAXWELL SOLVERSFOR ELECTROMAGNETIC PARTICLE-IN-CELL:

ALGORITHMS AND ADVANTAGES

Remi Lehe

LBNL

Invited TalkClassification: C-2, B-2, F-2

Electromagnetic Particle-In-Cell codes have been used to simu-late both radio-frequency accelerators and plasma-based acceler-ators. In this context, the Particle-In-Cell algorithm often usesthe finite-difference method in order to solve the Maxwell equa-tions. However, while this method is simple to implement andscales well to multiple processors, it is liable to a number of nu-merical artifacts that can be particularly serious for simulationsof accelerators. An alternative to the finite-difference method isthe use of spectral solvers, which are typically less prone to nu-merical artifacts. In this talk, I will review recent progress inthe use of spectral solvers for simulations of plasma-based accel-erators. This includes techniques to scale those solvers to largenumber of processors, extensions to cylindrical geometry, andadaptations to specific problems such as boosted-frame simula-tions.

——————————————–

ESS ACCELERATOR LATTICE DESIGN STUDIESAND AUTOMATIC SYNOPTIC DEPLOYMENT

Y. Levinsen, M. Eshraqi, T. Grandsaert, A. Jansson, H.Kocevar, O. Midttun, N. Milas, R. Miyamoto, C. Plostinar,

A. Ponton, R. de Prisco, T. Shea, H. D. Thomsen

European Spallation Source ERIC, Sweden & AarhusUniversity, Denmark

Classification: D-1

The European Spallation Source (ESS) is currently under con-struction in south of Sweden. A highly brilliant neutron sourcewith a 5 MW proton driver will provide state of the art experi-mental facilities for neutron science. A peak proton beam powerin the accelerator of 125 MW means that excellent control overthe beam losses becomes essential. The beam physics design of

34

the ESS accelerator is in a TraceWin format, for which we havedeveloped revision control setup, automated regression analysisand deployment of synoptic viewer and tabulated spreadsheets.This allows for an integrated representation of the data that arealways kept synchronized and available to other engineering dis-ciplines. The design of the accelerator lattice has gone throughseveral major and minor iterations which are all carefully anal-ysed. In this contribution we present the status of the latest stud-ies which is the first time a complete end-to-end study beginningfrom the ion source has been performed.

——————————————–

GENETIC ALGORITHM ENHANCED BYMACHINE LEARNING FOR DYNAMIC

APERTURE OPTIMIZATION

Yongjun Li, Weixing Cheng, Li Hua Yu, and RobertRainer

Brookhaven National Laboratory

Invited TalkClassification: E-2

With the aid of machine learning techniques, the genetic algo-rithm has been enhanced and applied to the multi-objective op-timization problem presented by the dynamic aperture of theNSLS-II Ring. During the evolution employed by the geneticalgorithm, the population is classified into different clusters. Theclusters with top average fitness are given elite status. Interven-tion is implemented by repopulating some potentially competi-tive candidates based on the accumulated data. These candidatesreplace randomly selected candidates among the original datapool. The average fitness of the population is improved whilediversity is not lost. The quality of the population increases andproduces more competitive descendants accelerating the evolu-tion process significantly. When identifying the distribution ofoptimal candidates, they appear to be located in isolated islandswithin the search space. Some of these optimal candidates havebeen experimentally confirmed at the NSLS-II storage ring. Themachine learning techniques that exploit the genetic algorithmcan also be used in other population-based optimization prob-lems such as particle swarm algorithm.

——————————————–

PYAOPT OPTIMIZATION SUITE AND ITSAPPLICATION ON ASTRA-SIMULATED SRF MEV

GUN DESIGN FOR UEMS

Ao Liu, Chunguang Jing

Euclid Techlabs

Classification: B-2

In order to achieve sharp, high resolution real-time imaging, elec-trons in a MeV UEM (ultrafast electron microscope) beamlineneed to minimize instabilities. The Superconducting RF (SRF)photocathode gun is a promising candidate to produce highly sta-ble electrons for UEM/UED applications. It operates in an ul-trahigh Q, CW mode, and dissipates a few watts of RF power,

which make it possible to achieve a 10s ppm level of beam sta-bility by using modern RF control techniques. In order to findthe best performance of the gun design, an optimization proce-dure is required. pyaopt is a Python-based optimization suitethat supports multi-objective optimizations using advanced algo-rithms. In this paper, the novel SRF photogun design and itsoptimizations through pyaopt and Astra’s beam simulations willbe discussed.

——————————————–

URGENT NEED OF START-TO-ENDSIMULATIONS FOR SHANGHAI CW HARD

X-RAY FEL PROJECT

Bo Liu, Qiang Gu, Zhen Wang, Meng Zhang, Si Chen,Haixiao Deng, Chao Feng, Dong Wang, Zhentang Zhao

Shanghai Institute of Applied Physics

Classification: A-1, B-1

Shanghai has started to construct the X-ray FEL facility SHINE(Shanghai high repetition rate XFEL and extreme light facility),which is based on a 8 GeV CW-SRF linac and will build threeundulator lines in the first stage. Designs of the gun, the injector,the linac, the distribution section and the FEL lines have alreadybeen done and will be presented here. Prelimilary study showsthat comprehensive study of the beam and FEL properties withstart-to-end simulations is really necessary.

——————————————–

BEAM ALIGNMENT SIMULATION ON THEBEAMLINE OF A PROTON THERAPY FACILITY

Xu Liu, Qushan Chen, Bin Qin, Guangyao Feng

Huazhong University of Science and Technology

Classification: D-1

Proton therapy is now recognized as one of the most effectiveradiation therapy methods for cancers. A proton therapy facilitywith multiple gantry treatment rooms is under development inHUST (Huazhong University of Science and Technology). Mis-alignments of magnets and beam diagnostics instruments inducethe offset of the beam trajectory, which will influence the clini-cal therapeutic effect. This paper describes the beam alignmentsimulations based on response matrix and this technology is ap-plied to the design of the HUST-PTF beamline. To perform thisstudy, we use the simulation code ELEGANT, and utilize theglobal correction method. By optimizing the layout of correctorsand beam position monitors, we completed the beam correctioncalculation. The results show that the accuracy of center beamtrajectory in the iso-center is better than 0.5 mm, meeting physi-cal and clinical requirements.

——————————————–

35

STATISTICAL ANALYSIS OF THE EIGENMODESPECTRUM IN THE SRF CAVITIES WITH

MECHANICAL IMPERFECTIONS

A. Lunin, T. Khabiboulline, N. Solyak, A. Sukhanov andV. Yakovlev

Fermilab

Classification: C-2

The SRF technology is progressing rapidly over last decadestoward high accelerating gradients and low surface resistancemaking feasible the particle accelerators operation with highbeam currents and long duty factors. However, the coherent RFlosses due to HOM radiation becomes a limiting factors for theseregimes. In spite of the operating mode, which is tuned sepa-rately, the parameters of HOMs vary from one cavity to anotherdue to finite mechanical tolerances during cavities fabrication. Itis vital to know in advance the spread of HOM parameters in or-der to predict unexpected cryogenic losses, overheating of beamline components and to keep stable beam dynamics. In this paperwe present the method of generating the unique cavity geometrywith imperfections while preserving operating mode frequencyand field flatness. Based on the eigenmode spectrum calculationof series of randomly generated cavities we can accumulate thedata for the evaluation the HOM statistics. Finally we describethe procedure for the estimation of the probability of the resonantHOM losses in the SRF resonators. The study of these effectsleads to specifications of SC cavity and cryomodule and can sig-nificantly impact on the efficiency and reliability of the machineoperation.

——————————————–

SIMULATION CHALLENGES FOR ERHICBEAM-BEAM STUDY

Y. Luo, Y. Hao, J. Qiang, Y. Roblin

Brookhaven National Laboratory

Classification: A-1, A-2, D-1

The 2015 Nuclear Science Advisory Committee Long Rang Planidentified the need for an electron-ion collider facility as a gu-lon microscope with capabilities beyond those of any existingaccelerator complex. To reach the required high energy, high lu-minosity, and high polarization, the eRHIC design based on theexisting heady ion and polarized proton collider RHIC adopts avery small beta-function at the interaction point, a high collisionrepetition rate, and a novel hadron cooling scheme. Collisionwith a full crossing angle of 22 mrad and crab cavities for bothelectron and proton rings are required. In this article, we willpresent the high priority R&D items related to beam-beam inter-action for the current eRHIC design, the simulation challenges,and our plans to address them.

——————————————–

SIMULATIONS OF COHERENT ELECTRONCOOLING WITH FREE ELECTRON LASER

AMPLIFIER AND PLASMA-CASCADEMICRO-BUNCHING AMPLIFIER

Jun Ma, Gang Wang, Vladimir N. Litvinenko

Brookhaven National Laboratory

Classification: D-1

SPACE is a parallel, relativistic 3D electromagnetic Particle-in-Cell (PIC) code used for simulations of beam dynamics and in-teractions. An electrostatic module has been developed with theimplementation of Adaptive Particle-in-Cloud method. Simula-tions performed by SPACE are capable of various beam distri-bution, different types of boundary conditions and flexible beamline, as well as sufficient data processing routines for data anal-ysis and visualization. Code SPACE has been used in the simu-lation studies of coherent electron cooling experiment based ontwo types of amplifiers, the free electron laser (FEL) amplifierand the plasma-cascade micro-bunching amplifier.

——————————————–

PLASMA WAKEFIELD START TO ENDACCELERATION SIMULATIONS, FROM

PHOTOCATHODE TO FEL WITH SIMULATEDDENSITY PROFILES

Alberto Marocchino

INFN Laboratori Nazionali di Frascati, Frascati, Italy

Invited TalkClassification: A-2

Plasma Wakefield acceleration is a promising new accelera-tion technique that profit by a charged bunch, e.g. an electronbunch, to break the neutrality of a plasma channel to producea wake where a trailing bunch is eventually accelerated. Thequest to achieve extreme gradient conserving high brightness hasprompted to a variety of new approaches and techniques. Mostof the proposed schemes are however limited to the only plasmachannel, assuming in the vast majority of cases, ideal scenarios(e.g. ideal bi-gaussian bunches and uniform density plasma chan-nels). Realistic start-to-end simulations from the photocathode toa FEL via a plasma accelerating section are a fundamental step tofurther investigate realistic scheme possibilities, the underlyingphysics, and future applications. To remove ideal simplifications,the SPARC LAB simulation team is simulating bunches from thephoto-cathode and tracking them all the way to the plasma. Sim-ilarly, the density profiles, where bunches evolve and accelerate,are calculated with a magneto-hydrodynamic code. The densityprofile is imported into the particle in cell codes used to calcu-late the particle evolution within the plasma section. The use ofa multitude of codes, involving different architectures, physicalunits, and programming languages, made necessary the defini-tion of code interfacing and pipe-processes to ensure a properpipeline of tools that are traditionally used in different fields aredo not often come across. By combining the different numericalcodes (particle tracker, particle in cell, magneto-hydrodynamicsand FEL codes) we could propose a first realistic start-to-end

36

simulation from the photo-cathode to a FEL lasering for a possi-ble upcoming Italian PWFA-FEL facility. Such a work is con-ducted with a great focus on code reliability and data repro-ducibility. The Italian PWFA experimental team uses a capil-lary to control and tailor the plasma density profile, we couldperform preliminary code comparison and validation against ex-perimental data. Code validation has also been possible for pas-sive plasma lens experiments, where the detailed experimentalsix-dimensional phase space reconstruction had allowed a directcomparison with the numerical tools.

——————————————–

EFFICIENT MODELING OF LASER WAKEFIELDACCELERATION THROUGH THE PIC CODE

SMILEI IN CILEX PROJECT

Francesco Massimo (1), Arnaud Beck (1), JulienDerouillat (2), Mickael Grech (3), Frederic Perez (3),

Imen Zemzemi (1), Arnd Specka (1)

(1) Laboratoire Leprince-Ringuet - Ecole Polytechnique,CNRS-IN2P3, Palaiseau 91128, France - (2) Maison de laSimulation, CEA, CNRS, Universite Paris-Sud, UVSQ,

Universite Paris-Saclay, F-91191 Gif-sur-Yvette, France -(3) Laboratoire d’Utilisation des Lasers Intenses, CNRS,

Ecole Polytechnique, CEA, Universite Paris-Saclay,UPMC Universite Paris 06: Sorbonne Universites,

F-91128 Palaiseau Cedex, France

Classification: B-2, F-1

The design of plasma acceleration facilities requires considerablesimulation effort for each part of the machine, from the plasmainjector and/or accelerator stage(s), to the beam transport stage,from which the accelerated beams will be brought to the usersor possibly to another plasma stage. The urgent issues and chal-lenges in simulation of multi-stage acceleration with the Apollonlaser of CILEX facility will be addressed. To simulate the beaminjection in the second plasma stage, additional physical mod-els have been introduced and tested in the open source Particlein Cell collaborative code Smilei. The efficient initialisation ofarbitrary relativistic particle beam distributions through a Pythoninterface allowing code coupling and the self consistent initialisa-tion of their electromagnetic fields will be presented. The com-parison between a full PIC simulation and a simulation with arecently developed envelope model, which allows to drasticallyreduce the computational time, will be also shown for a test caseof laser wakefield acceleration of an externally injected electronbeam.

——————————————–

DESIGN OF A COMPACT PERMANENT MAGNETSPECTROMETER FOR CILEX/APOLLON

M. Khojoyan (1), J. Prudent (1), A. Cauchois (1), F.Massimo (1) and A. Specka (1)

(1) Laboratoire Leprince-Ringuet - Ecole Polytechnique,CNRS-IN2P3, Palaiseau 91128, France

Classification: C-1, C-2, D-1

Laser wakefield acceleration experiments make extensive use ofsmall permanent magnets or magnet assemblies for diagnosticand focusing of electron beams produced in plasma accelera-tors. This choice is motivated by the ease of operation insidevacuum chambers, absence of power-supplies and feedthroughs,and potentially lower cost. Indeed, in these experiments space isat premium, and compactness is frequently required. At the sametime, these magnets need to have a large angular acceptance forthe divergent electron beams, which imposes constraints on thegap size. We will present the optimized design and character-ization of a 100 mm long, 2.1 Tesla permanent magnet dipole.Furthermore, we will present the implementation of this mag-net in a spectrometer that will measure the energy spectrum ofelectrons of [60-2000] MeV with a few percent resolution in theCILEX/APOLLON 10PW laser facility in France.

——————————————–

LIGHTSOURCE UNIFIED MODELINGENVIRONMENT (LUME), A START-TO-END

SIMULATION FRAMEWORK FOR ELECTRONSAND PHOTONS.

Christopher Mayes, Paul Fuoss, Chuck Yoon

SLAC National Accelerator Laboratory

Invited TalkClassification: A-2, D-1

Since first light at LCLS, there has been continuous inventionof new operating modes, introduction of new optical elements,and apid improvement in detectors. While these improvementshave led to new experiments with much greater scientific im-pacts, their transfer to user operations has often taken several ex-perimental runs (many months to years). The integration of thesetechnical advances into scientific programs would be greatly ac-celerated by a modeling tool that allowed for quantitative assess-ment of the impact on scientific programs of facility improve-ments. To this end, SLAC is developing the Lightsource UnifiedModeling Environment (LUME) for unified modeling of X-rayfree electron laser (XFEL) performance. This modeling tool willbe built in several stages with an initial focus on quantitativeprediction of critical parameters of the X-ray pulses deliveredto experimental stations. This initial development will be fol-lowed by incorporation of X-ray-sample interaction and detectorperformance. This project will take a holistic approach startingwith the simulation of the electron beams, to the production ofthe photon pulses and their transport through the optical com-ponents of the beamline, their interaction with the samples andthe simulation of the detectors, followed by the analysis of simu-lated data. LUME will leverage existing, well-established codes[Astra, Bmad, Elegant, Genesis, Impact for electrons, Genesis

37

1.3 for FEL simulation, and the “Synchrotron Radiation Work-shop” (SRW) for X-ray optics] that will be driven and configuredby a coherent high-level framework. The high-level frameworkwill build on the Simex platform being developed by the Euro-pean Cluster of Advanced Laser Light Sources (EUCALL). Theplatform will be built with an open, well-documented architec-ture so that science groups around the world can contribute spe-cific experimental designs and software modules, advancing boththeir scientific interests and a broader knowledge of the opportu-nities provided by the exceptional capabilities of X-ray FELs.LUME will be the first platform in the world for unified mod-eling of XFEL performance. LUME’s optimization capabilitieswill guide SLAC accelerator physicists in developing world lead-ing XFEL performance. LUME will identify performance bottle-necks, both in the accelerator and photon transport, and enhanceoperational efficiency and reliability. The complete integrationof electron and X-ray processes will allow LCLS scientists to in-vent instruments that optimally use those unique X-ray beams.Finally and most importantly, the ability to simulate experimentswill stimulate the development of new approaches to the scien-tific and technological challenges facing the country, maximizingthe impact of DoE’s investment in cutting-edge X-ray sources.

——————————————–

A FULL FIELD-MAP MODELING OFCORNELL-BNL CBETA 4-PASS ENERGY

RECOVERY LINAC

F. Meot, S. Brooks, J. Crittenden, D. Trbojevic, N.Tsoupas

BNL and Cornell

Classification: A-2

The Cornell-BNL Electron Test Accelerator (CBETA) is a four-pass, 150 MeV energy recovery linac (ERL), now in constructionat Cornell. A single fixed-field alternating gradient (FFAG) beamline recirculates the four energies, 42, 78, 114 and 150 MeV.The return loop is comprised of 107 quadrupole-doublet cells,built using Halbach permanent magnet technology. Spreader andcombiner sections (4 independent beam lines each) connect the36 MeV linac to the FFAG loop. We present here a start-to-endsimulation of the 4-pass ERL, entirely, and exclusively, based onthe use of magnetic field maps to model the magnets and correc-tors. There are paramount reasons for that and this is discussed,detailed outcomes are presented, together with comparisons withregular beam transport (mapping based) techniques.

——————————————–

POLARIZATION LIFETIME IN AN ELECTRONSTORAGE RING, AN ERGODIC APPROACH IN

ERHIC EIC

Francois Meot

Brookhaven National Laboratory

Classification: D-2

Electron polarization in a storage ring is subject to two verylong term effects: Sokolov-Ternov polarization and depolariza-tion by diffusion. This leads to an equilibrium state, over a verylong time scale, and, simulation-wise, is highly CPU time andmemory consuming. Simulations aimed at determining optimalring storage energy in an electron-ion collider are always track-ing bunches with thousands of particles, and in addition for tooshort time scales due to HPC limitations. Based on considera-tions of ergodicity of electron bunch dynamics in the presenceof synchrotron radiation, and on the very slow depolarizationaimed at in a collider, tracking a single particle instead is inves-tigated. This saves a factor of more than 2 orders of magnitudesin the parameter CPU-time x Memory-allocation, it allows muchlonger tracking and thus accuracy on the evaluation of polariza-tion and time constants. The concept is illustrated with polar-ization lifetime and equilibrium polarization simulations at theeRHIC electron-ion collider.

——————————————–

SPACE CHARGE AND TRANSVERSEINSTABILITIES AT THE CERN SPS AND LHC

Elias Metral

CERN

Invited TalkClassification: D-1

At the CERN accelerator complex, only the highest energy ma-chine in the sequence, the LHC, with space charge parameterclose to one, sees a beneficial effect of space charge on trans-verse coherent instabilities. In the other circular machines of theLHC injector chain (PSB, PS and SPS), where the space chargeparameter is much bigger than one, space charge does not seemto play a major role. All the measurements and simulations per-formed so far in both the SPS and LHC will be reviewed andanalyzed in detail.

——————————————–

38

ANALYSIS OF EMITTANCE GROWTH DUE TOCOLLISIONAL PARTICLE NOISE IN A GRIDLESS

SPECTRAL POISSON SOLVER FOR FULLYSYMPLECTIC MULTIPARTICLE TRACKING

Chad Mitchell, Ji Qiang

Lawrence Berkeley National Laboratory

Classification: D-1, D-2, F-1

Gridless spectral methods for self-consistent space charge mod-eling possess several advantages over traditional momentum-conserving particle-in-cell methods, including the absence of nu-merical grid heating and the presence of an underlying multi-particle Hamiltonian. Nevertheless, evidence of collisional parti-cle noise remains. For a class of such 2D algorithms, we provideanalytical models of the numerical field error, the optimal choiceof spectral mode cutoff, and the numerical emittance growth pertimestep. We compare these results with the emittance growthmodels of Struckmeier, Hoffman, Kesting, and others.

——————————————–

BEAMLINE MAP COMPUTATION FORPARAXIAL OPTICS

Boaz Nash

RadiaSoft LLC

Classification: B-1, C-2

Modeling of radiation transport is an important topic tightly cou-pled to many charged particle dynamics simulations for syn-chrotron light sources and FEL facilities. The radiation is deter-mined by the electron beam and magnetic field source, and thenpasses through beamlines with focusing elements, apertures andmonochromators, in which one may typically apply the paraxialapproximation of small angular deviations from the optical axis.The radiation is then used in a wide range of spectroscopic ex-periments, or else may be recirculated back to the electron beamsource, in the case of an FEL oscillator. The Wigner functionrepresentation of electromagnetic wavefronts has been describedin the literature and allows a phase space description of the ra-diation, similar to that used in charged particle dynamics. It canencompass both fully and partially coherent cases, as well as po-larization. Here, we describe the calculation of a beamline mapthat can be applied to the radiation Wigner function, reducing thecomputation time. We discuss the use of ray tracing and waveoptics codes for the map computation and benchmarking. Weconstruct a four crystal 1:1 imaging beamline that could be usedfor recirculation in an XFEL oscillator, and benchmark the mapbased results with SRW wavefront simulations.

——————————————–

COMPARISON OF MODEL BASED ANDHEURISTIC OPTIMIZATION ALGORITHMS

APPLIED TO PHOTOINJECTORS USINGLIBENSEMBLE

Nicole Neveu, Jeffrey Larson, Stephen Hudson, LindaSpentzouris

Illinois Institute of Technology, Argonne NationalLaboratory

Classification: E-2, F-2

Genetic algorithms are common and often used in the acceler-ator community. They require large amounts of computationalresources and empirical adjustment of hyperparameters. Modelbased methods are significantly more efficient, but often labeledas unreliable for the nonlinear or unsmooth problems that can befound in accelerator physics. We investigate the behavior of bothapproaches using a photoinjector operated in the space chargedominated regime. All optimization runs are coordinated andmanaged by the Python library libEnsemble, which is developedat Argonne National Laboratory.

——————————————–

CHALLENGES IN SIMULATING BEAMDYNAMICS OF DIELECTRIC LASER

ACCELERATION

Uwe Niedermayer

TEMF TU Darmstadt

Plenary TalkClassification: B-2, D-1

Dielectric Laser Acceleration (DLA) achieves highest gradientsin non-plasma accelerators by using the inverse Smith-Purcelleffect on a dielectric grating, which is almost lossless at the re-spective laser wavelength. The use of dielectrics increases thebreakdown field strength, and thus the achievable gradient, bya factor of at least 10 in comparison to metals. Experimentalbreakthroughs in DLA led to the Accelerator on a Chip (ACHIP)project, funded by the Gordon and Betty Moore Foundation from2015 till 2020. In ACHIP, our main goal is to build an accelera-tor on a silicon chip, which can accelerate electrons from below100keV to above 1MeV with a gradient of at least 100MeV/m.For stable acceleration on the chip, magnet-only focusing tech-niques are insufficient to compensate the strong acceleration de-focusing. Thus higher spatial harmonic and Alternating PhaseFocusing (APF) laser based focusing techniques have been de-veloped. We have also developed the simplified symplectic track-ing code DLAtrack6D, which makes use of the periodicity andapplies only one kick per DLA cell, which is calculated by theFourier coefficient of the synchronous spatial harmonic. Dueto the coupling of the cells, the Fourier coefficients are not flatbut a field flatness optimization (similarly as in multi-cell cavi-ties) needs to be performed. The effect of the APF-drifts and theend cells need to be studied and mitigated by individual design.Moreover, fabrication tolerances and misalignment need investi-gation and mitigation by improving the beam dynamics robust-ness. The simulation of the entire accelerator on a chip by a PICcode is possible, but not practical for optimization purposes since

39

a cluster computer is already required for a single run. Finally,we also outline the treatment of wake field effects at attosecondbunches in the grating by DLAtrack6D, where the wake field iscomputed by an external solver.

——————————————–

PARTICLE-IN-CELL SIMULATION OF ABUNCHED ELECTRONS BEAM ACCELERATIONIN A TE113 CYLINDRICAL CAVITY AFFECTEDBY A STATIC INHOMOGENEOUS MAGNETIC

FIELD

Eduardo A. Orozco (1), Victor E. Vergara (2), Jesus D.Gonzalez (2) and Jesus R. Beltran (2)

(1) Universidad Industrial de Santander, Bucaramanga,Colombia; (2) Universidad del Magdalena, Santa Marta,

Colombia.

Classification: B-2, C-2, D-1

The results of the relativistic full electromagnetic Particle-in-cell(PIC) simulation of a bunched electrons beam accelerated in acylindrical cavity mode TE113 in the presence of a static inho-mogeneous magnetic field are presented. This type of acceler-ation is known as Spatial AutoResonance Acceleration (SARA)[1]. The magnetic field profile is such that it keeps the electronsbeam in the acceleration regime along their trajectories. Numer-ical experiments of bunched electrons beam with the concentra-tions in the range 108-1011 cm−3 in a linear TE113 cylindri-cal microwave field of a frequency of 2.45GHz and an ampli-tude of about 14kV /cm show that it is possible accelerate thebunched electrons up to energies of the order of 300keV with-out serious defocalization effect. A comparison between thedata obtained from the full electromagnetic PIC simulations andthe results derived from the relativistic Newton-Lorentz equationin a single particle approximation [2] is carried out. This ac-celeration scheme can be used to produce hard x-ray [3]. [1]Dugar-Zhabon, V. D., & Orozco, E. A. (2009). Cyclotron spa-tial autoresonance acceleration model. Physical Review SpecialTopics-Accelerators and Beams, 12(4), 041301. [2] Vergara, V.E., Gonzalez, J. D., Beltran, J. R., & Orozco, E. A. (2017, De-cember). Electrons acceleration in a TE113 cylindrical cavityaffected by a static inhomogeneous magnetic field. In Journal ofPhysics: Conference Series (Vol. 935, No. 1, p. 012076). IOPPublishing. [3] Dugar-Zhabon, V. D., & Orozco, E. A. (2017).Compact self-resonant x ray source. (USA Patent: 9,666,403 )

——————————————–

COMPUTATIONAL BEAM DYNAMICSREQUIREMENTS FOR FRIB

Peter Ostroumov

MSU, FRIB

Plenary TalkClassification: A-1, D-1, E-2

The Facility for Rare Isotope Beams (FRIB) being built at Michi-gan State University moved to the commissioned stage in thesummer of 2017. There were extensive beam dynamics simula-tions in the FRIB driver linac during the design stage. Recently,we have used TRACK and IMPACT simulation codes to studydynamics of ion beam contaminants extracted from the ECR to-gether with main ion beam. The contaminant ion species canproduce significant losses after the stripping. These studies re-sulted in development of beam collimation system at relativelylow energy of 16 MeV/u and room temperature bunchers in-stead of originally planned SC cavities. Commissioning of theFront End and the first 3 cryomodules enabled detailed beam dy-namics studies experimentally which were accompanied with thesimulations using above-mentioned beam dynamics codes andoptimization code FLAME. There are significant challenges inunderstanding of beam dynamics in the FRIB linac. The mostcomputational challenges are in the following areas: (1) Simula-tion of the ion beam formation and extraction from the ECR; (2)Development of the virtual accelerator model available on-lineboth for optimization and multi-particle simulations. The vir-tual model should include realistic accelerator parameters includ-ing device misalignments; (3) Large scale simulations to supporthigh-power ramp up of the linac with minimized beam losses; (4)Interaction of the beam with the gas stripper which is the backupoption for high power operation of the linac. Work supported bythe U.S. Department of Energy Office of Science under Coopera-tive Agreement DE-SC0000661 and the National Science Foun-dation under Cooperative Agreement PHY-1102511, the State ofMichigan and Michigan State University.

——————————————–

STUDY OF ELECTRON CYCLOTRONRESONANCE ACCELERATION BY

CYLINDRICAL TE011 MODE

Oswaldo Otero Olarte(1), Eduardo Alberto Orozco (1),Ana Marıa Herrera (2)

(1) Universidad Industrial de Santander, Bucaramanga,Colombia, (2) Heidelberg Institute for Theoretical Studies,

Heidelberg, Germany

Classification: B-2

In this work, we present results from analytical and numericalstudies of the electron acceleration by a TE011 cylindrical mi-crowave mode in a static homogeneous magnetic field underelectron cyclotron resonance (ECR) condition. The stability ofthe orbits is analyzed using the particle orbit theory. In order toget a better understanding of the interaction wave-particle we de-compose the azimuthally electric field component as the superpo-sition of right and left hand circular polarization standing waves.The trajectory, energy and phase-shift of the electron are found

40

through a numerical solution of the relativistic Newton-Lorentzequation in a finite difference method by the Boris method. It isshown that an electron longitudinally injected with an energy of7 keV in a radial position r=Rc/2, being Rc the cavity radius, isaccelerated up to energy of 90 keV by an electric field strengthof 14 kV/cm and frequency of 2.45 GHz. This energy can beused to produce X-ray for medical imaging. These results can beused as a starting point for the study the acceleration of electronsin a magnetic field changing slowly in time (GYRAC), whichhas some important applications as the electron cyclotron reso-nance Ion proton accelerator (ECR-IPAC) for cancer therapy andto control plasma bunches with relativistic electrons.

——————————————–

MULTIPASS SIMULATIONS OF SPACE CHARGECOMPENSATION USING ELECTRON COLUMNS

AT IOTA RING

Chong Shik Park, Moses Chung, Ben Freemire, ChadMitchell, Greg Penn, Giulio Stancari, Eric Stern

Korea University, UNIST, NIU, LBNL, Fermilab

Classification: D-1

Defocusing repulsive forces due to self space charge fields leadsto degradation of high-intensity particle beams. Being of par-ticular concern for low- and medium-energy proton beams, theyresult in the emittance growth, beam halo formation, and beamlosses. They set stringent limits on the intensity frontier ac-celerators, therefore, the mitigation of space charge effects isa crucial challenge to improve the proton beam intensity. Thespace charge effects in the positively charged proton beams canbe effectively compensated by using negatively charged elec-tron columns. In this paper, we present the key parameters ofthe electron columns for the space charge compensation (SCC)and discuss results of the Warp3D numerical simulations for thematching of the transverse and longitudinal charge distributionsof electrons produced by the high intensity proton beam and ac-cumulated in the column. In order to investigate the evolution ofboth the electron column and the proton beam over multi-passes,we track the proton beam further within in the IOTA ring usingSynergia and IMPACT.

——————————————–

SYMPLECTIC PARTICLE-IN-CELL

Thomas Planche

TRIUMF

Invited TalkClassification: D-1, D-2

This is a review talk on symplectic self-consistent algorithms forthe study of space-charge effects in particle accelerators. Startingfrom the Low Lagrangian for collision-less plasmas, I will showhow to derive a Hamiltonian for relativistic beams in particle ac-celerators. From this Hamiltonian one can derive the evolution ofthe particle distribution as well as the self-field. Having obtaineda discretized version of this Hamiltonian, I will discuss the use of

map methods to achieve self-consistent symplectic multi-particletracking.

——————————————–

LONGITUDINAL BEAM DYNAMICS IN FRIB ANDREA LINACS

A.S. Plastun, P.N. Ostroumov, A.C.C. Villari, Q. Zhao

Facility for Rare Isotope Beams, Michigan StateUniversity, 48824, East Lansing, MI, USA

Classification: E-2

The Front-End and first three cryomodules of the Facility forRare Isotope Beam (FRIB) at Michigan State University (MSU)has been commissioned in July, 2018. The paper describes pro-cedures developed for the online tuning of the longitudinal beamdynamics through the FRIB linac. These procedures include(a) an automated simulation-based tuning of the multi-harmonicbuncher, (b) measurements and simulations of the RFQ thresh-old voltage and longitudinal acceptance, (c) RF phase scans ofthe rebunchers and superconducting accelerating cavities. WhileFRIB is being commissioned, the reaccelerator (ReA3) for rareisotope beams (RIBs) is being upgraded. In order to matchany ReA3 beam both to the following upgrade cryomodules andphysics experiments’ requirements, room temperature rebunch-ers/debunchers are being designed. The design procedure in-cludes the electromagnetic, thermal and mechanical simulationsand optimizations.

——————————————–

MAGNETIZED ELECTRON COOLINGSIMULATIONS FOR JLEIC

Ilya Pogorelov (1), David Bruhwiler (1), Christopher Hall(1), Stephen Webb (1), Dan Abell (1), Yury Eidelman (1),Johan Carlsson (1), James Gerity (2), Peter McIntyre (2)

(1) RadiaSoft LLC, (2) Texas A&M U.

Classification: D-1, D-2, F-1, A-2

Relativistic magnetized electron cooling in untested parameterregimes is essential to achieve the ion luminosity requirementsof proposed electron-ion collider (EIC) designs. Therefore, ac-curate calculations of magnetized dynamic friction are required,with the ability to include all relevant physics that might increasethe cooling time, including space charge forces, field errors andcomplicated phase space distributions of imperfectly magnetizedelectron beams. We present simulations relevant to the JLEIC de-sign, using the BETACOOL and JSPEC codes. We also presentrecent work on Warp simulations of the electron beam throughthe solenoid field. Space charge neutralization is provided byimpact ionization of a background hydrogen gas. For optimalcooling it is essential that space charge be sufficiently neutral-ized. We also present recent work on a new analytic treatment ofmomentum transfer from a single magnetized electron to a drift-ing ion, and its use for calculations of dynamic friction. Thiswork is supported by the U.S. DOE Office of Science, Office ofNuclear Physics, under Award Number DE-SC0015212.

41

——————————————–

EFFICIENT COMPUTATION OF LOSSY HIGHERORDER MODES IN COMPLEX SRF CAVITIES

USING REDUCED ORDER MODELS ANDNONLINEAR EIGENVALUE PROBLEM

ALGORITHMS

Hermann W. Pommerenke, Johann D. Heller, Ursula vanRienen

Institute of General Electrical Engineering, University ofRostock, Germany

Invited TalkClassification: C-2

Superconducting radio frequency (SRF) cavities meet the de-manding performance requirements of modern accelerators andhigh-brilliance light sources. For the operation and design ofsuch resonators, a very precise knowledge of their electromag-netic resonances is required. The non-trivial cavity shape de-mands a numerical solution of Maxwell’s equations to computethe resonant eigenfrequencies, eigenmodes, and their losses. Forlarge and complex structures this is hardly possible on conven-tional hardware due to the high number of degrees of freedomrequired to obtain an accurate solution. Here, we propose amethod which can solve the considered problems on worksta-tion computers without extensive simplification of the structureitself. First, the State-Space Concatenation scheme (SSC) is ap-plied to the complex, closed and thus lossless RF structure. SSCemploys a combination of model order reduction and domain de-composition, greatly reducing the computational effort by effec-tively constraining the considered frequency domain. Next, aperturbation approach based on SSC is used to describe the res-onances of the same geometry subject to external losses. Dueto the boundary conditions this results in a nonlinear eigenvalueproblem (NLEVP). The NLEVP can be solved efficiently byNewton’s method, or in combination with a contour integral al-gorithm. We present the general workflow to compute the elec-tromagnetic resonances with an emphasis given to the algorithmused to solve the arising NLEVP.

——————————————–

ADVANCES IN SIMULATION OF HIGHBRIGHTNESS/HIGH INTENSITY BEAMS

Ji Qiang

LBNL

Plenary TalkClassification: D-1

Large-scale advanced modeling of high intensity/high brightnessbeams plays an important role in beam dynamics study and ac-celerator design. In this paper, we report on recent progress instart-to-end simulation of high brightness electron beam in x-rayFEL accelerator and progress in long-term tracking simulation ofspace-charge effects in high intensity proton beam.

——————————————–

APPROACHES TO OPTIMIZING SPINTRANSMISSION IN LATTICE DESIGN

Vahid Ranjbar

BNL

Classification: D-2, E-2, A-2

We present our experiences in optimizing the proposed RapidCycling Synchrotron (RCS) injector for the eRHIC Storage ringand the RHIC 2017 lattice. We have develop python code to drivelattice calculations in MADX which are then used to calculatespin resonances using the DEPOL algorithm. This approach hasbeen used to minimize intrinsic spin resonances during the RCSacceleration cycle while controlling lattice parameters such asdispersion and beta functions. This approach has also been usedto construct localized imperfection bumps using a spin responsematrix and SVD. This approach has also been used to reduce in-terfering intrinsic spin resonances during the RHIC accelerationramp.

——————————————–

MACHINE LEARNING FOR X-RAYFREE-ELECTRON LASERS

Daniel Ratner

SLAC

Invited TalkClassification: B-1, E-2, F-2

X-ray Free Electron Lasers (XFELs) are among the most com-plex accelerator projects in the world today. With large param-eter spaces, sensitive dependence on beam quality, huge datarates, and challenging machine protection, there are expandingopportunities to apply machine learning (ML) to XFEL opera-tion. In this talk I will summarize some promising ML methodsfor XFELs, and highlight recent examples of successful applica-tions.

——————————————–

SIMULATIONS OF LONGITUDINAL BEAMSTABILITY IN THE CERN SPS WITH BLOND

J. Repond, K. Iliakis, I. Karpov, A. Lasheen, D. Quartulo,M. Schwarz, E. Shaposhnikova, H. Timko

CERN

Classification: D-1, F-1

The Super Proton Synchrotron (SPS) at CERN, the Large HadronCollider (LHC) injector, is currently pushed to its limits for theproduction of the LHC proton beam while beam quality and sta-bility in the longitudinal plane are influenced by many effects.Particle simulation codes become an essential tool to study thebeam instabilities. BLonD, developed at CERN, is a 2D particle-tracking simulation code, modeling the phase space of singleand multi-bunch beams in multi-harmonic RF systems. Com-putation of collective effects due to the machine impedance andspace charge is available on a multi-turn basis. Various beam

42

control loops of the RF system are implemented (phase, fre-quency and synchro-loops, and one-turn delay feedback) as wellas RF phase noise injection used for controlled emittance blow-up. The challenges of the longitudinal beam stability simula-tions during long SPS acceleration cycle (19.93 s) are the varietyof effects impacting beam dynamics (beam loading, instabilities,particle losses, controlled blow-up, double RF system operation,low-level RF control, bunch distribution, etc.), the complicatedSPS impedance model containing broad and narrow-band reso-nant modes between 50 MHz and 4 GHz, and the large numberof bunches in the nominal LHC batch (288). This paper presentsa selection of BLonD simulation studies addressing these chal-lenges and with results substantiated, when possible, by beammeasurement data.

——————————————–

OPTIMIZATION STUDIES FOR THE K12 BEAMLINE AT THE CERN NORTH AREA

M. Rosenthal (1), D. Banerjee (1), J. Bernhard (1), M.Brugger (1), N. Charitonidis (1), B. Dobrich (1), L.

Gatignon (1), A. Gerbershagen (1), E. Montbarbon (1), B.Rae (1), T. Spadaro (2), M. van Dijk (1)

(1) CERN, 1211 Geneva 23, Switzerland, (2) LaboratoriNazionali di Frascati, I-00044 Frascati, Italy

Classification: A-2, F-1, A-1

The North Area at the Super Proton Synchrotron at CERN has along history of fixed target experiments and R&D studies. The400 GeV/c proton beam is extracted from the SPS and guided totwo experimental halls (EHN1, EHN2) and an underground cav-ern (ECN3) located at the CERN Prevessin site. Currently, ECN3hosts the NA62 experiment studying rare decays of positivelycharged kaons into pions and two neutrinos. The required high-intensity kaon beam is provided by a new secondary beam line,designated K12, which was constructed in 2012. At its start, theprimary proton beam impinges on a beryllium target (T10) with anominal intensity of 3E12 protons per burst. The momentum se-lection is performed by a massive dump collimator (TAX), whichis surrounded by four bending magnets in an achromat configu-ration. A future proposal for NA62 within the Physics BeyondColliders (PBC) framework suggests the search for dark sectorparticles such as heavy neutral leptons, dark photons and axions.For this purpose, the T10 target will be moved out, dumping theprimary proton beam on the 3.2 m long TAX. Muons originat-ing in these interactions are a severe background for this kindof experiment, demanding an effective magnetic sweeping alongthe beam line. The simulation of production and transport ofthis muon background is computationally highly expensive andrequires precise magnetic field information of the entire beamline. Monte Carlo studies based on the program G4Beamlinecombined with analytical parametrisations are used to reduce thecomputational demands. In this contribution, benchmarking re-sults with already recorded data as well as results from the opti-mization studies will be presented.

——————————————–

NONLINEAR OPTICS AT UMER

Kiersten Ruisard, Brian Beaudoin, Irving Haber,Timothy Koeth

ORNL, University of Maryland

Invited TalkClassification: D-1

Design of accelerator lattices with nonlinear optics to suppresstransverse resonances is a novel approach and may be cru-cial for enabling low-loss high-intensity beam transport. Largeamplitude-dependent tune spreads, driven by nonlinear field in-serts, damp resonant response to driving terms. This presenta-tion will focus on simulations of the UMER lattice operated asa quasi-integrable system (1 invariant of transverse motion) witha single strong octupole insert. We will discuss the evolution ofsimulation models, including the observation of losses associatedwith the original operating point near a fourth-order resonance.Other operating points farther from this resonance are consideredand shown to be more promising.

——————————————–

CHALLENGES IN EXTRACTINGPSEUDO-MULTIPOLES FROM MAGNETIC

MEASUREMENTS

Stephan Russenschuck, Gianni Caiafa, Melvin Liebsch,Carlo Petrone

CERN, 1211 Geneva 23, Switzerland

Classification: C-1

Extracting the coefficients of the Fourier-Bessel series (knownas pseudo-multipoles or generalized gradients) from magneticmeasurements of accelerator magnets bears some technical andmathematical challenges. A novel design of a short, rotating-coilmagnetometer is required that does not intercept any longitudi-nal field components. Moreover, displacing short magnetometersstep-by-step along the axis of the magnet, delivers a signals forthe transversal multipoles that are convolutions of the multipolesand the sensitivity of the induction coil. The deconvolution ofthe measured signals has then to deal with the (noisy) measure-ment data from the magnetometer. Moreover, the compensationschemes for the main component, as implemented in long coilsused for the integrated harmonics, cannot be applied to the shortmagnetometers. The paper presents the theory of the data acqui-sition and deconvolution, as well as the design and productionof a rotating-coil magnetometer that consists of four layers offlexible printed circuit mounted on a precision machined shaft.The design aimed at maximizing the sensitivity factors for fieldharmonics up to order 13 and at a compensation (bucking) ratiofor the main component in the same range of what is achievablewith standard rotating coils. The design , the uncertainty analysis(yielding the manufacturing tolerances), the manufacturing chal-lenges, and the results of dipole and quadrupole field scans willbe presented. Ref: [1] Berz, M.: Modern Map Methods in Par-ticle Beam Physics, Academic Press, 1999. [2] Russenschuck,S.: Rotating- and translating-coil magnetometers for extractingpseudo-multipoles in accelerator magnets, COMPEL - The in-ternational journal for computation and mathematics in electrical

43

and electronic engineering, Vol. 36 Issue: 5, 2017 [3] Arpaia,P., Buzio, M., De Matteis, E., Russenschuck, S.: A rotating coiltransducer for magnetic field mapping, Jinst, 2015.

——————————————–

COMPUTATIONAL ACCELERATOR PHYSICS: ONTHE ROAD TO EXASCALE

Robert D Ryne

Lawrence Berkeley National Laboratory

Plenary TalkClassification: F-2

The first conference in what would become the ICAP series washeld in 1988. At that time the most powerful computer in theworld was a Cray YMP with 8 processors and a peak perfor-mance of about 2 gigaflops. Today the fastest computer in theworld has more than 2 million cores and a theoretical peak per-formance of nearly 200 petaflops. Compared to 1988, perfor-mance has increased by a factor of 100 million, accompaniedby huge advances in memory, networking, big data managementand analytics, etc. By the time of the next ICAP conference in2021 we can expect to be living with the first exascale comput-ers. In this talk I will describe the advances that have taken placein computational accelerator physics since this conference seriesbegan, with emphasis on current examples ranging from 1000’sof cores up to the petascale, and describe what to expect in theexascale regime of the future.

——————————————–

REPTIL - A RELATIVISTIC 3D SPACE CHARGEPARTICLE TRACKING CODE BASED ON THE

FAST MULTIPOLE METHOD

Steffen Schmid, Erion Gjonaj, Herbert De Gersem

TU Darmstadt, TEMF

Classification: D-1

Modern free electron lasers and high current energy recoverylinacs accelerate electron beams with particle bunch chargesreaching up to several nanocoulombs. Especially in the low en-ergy sections, such as the photoinjector of the accelerator, spacecharge interaction forces are the dominating effect influencingthe dynamics of the electron beam. A direct computation ofspace charge forces is numerically very expensive. Commonlyused simulation codes typically apply mesh based particle-in-cell methods (PIC) to solve this problem. Our simulation tool,REPTIL, is a relativistic, three-dimensional space charge track-ing code, which computes the interaction forces based on a mesh-less fast multipole method (FMM). The FMM based space chargesolver is more flexible regarding the choice of the interactionmodel and yields maximum accuracy for the near field forces be-tween particles. For this reason, the FMM is very suitable for thesimulation of the influence of space charge on the particle emis-sion process in high current photoinjectors. In this contribution,we present a numerical study of the efficiency and the accuracyof the method. Therefore, we perform a case study for the PITZ

photoinjector used for the European XFEL at DESY. Further-more, we compare the performance of REPTIL with commonlyused PIC codes like e.g. ASTRA. Finally, we investigate a hy-brid approach by using the FMM on a mesh. The latter methodmakes further increases in the particle number possible, whichtranslates to a higher resolution in the phase space of the electronbunch.

——————————————–

UPDATE ON THE STATUS OF LINAC PART OFTHE PYORBIT CODE

Andrei Shishlo

Oak Ridge National Lab

Classification: F-1

The structure and capabilities of the linac beam dynamics partof the PyORBIT code are presented. The PyORBIT is an opensource code, a descendant of the original ORBIT code that wasdeveloped at the Spallation Neutron Source (SNS) for design,commissioning, and studies of the ring. The linac part wasstarted 8 years ago to utilize PyORBIT classes and infrastructurefor the SNS linac simulations. The PyORBIT linac model hasits own lattice description that is necessary to include lattice el-ements significantly different from the PyORBIT ring elements.The most important among them are accelerating RF structures.The five different RF gap models recently implemented in Py-ORBIT are discussed. Some benchmarks of the PyORBIT withParmila, the XAL Online Model, and TraceWin code are pre-sented.

——————————————–

MUON G-2: AN INTERPLAY OF BEAMDYNAMICS AND HEP

Michael Syphers

Northern Illinois University

Classification: A-2, D-1

The Fermilab experiment E989, Muon g-2, unites particle beamphysics with a high energy physics experiment in a unique way.The close interplay of the understanding of particle beam dynam-ics and the preparation of the beam properties with the experi-mental measurement is tantamount to the reduction of systematicerrors in the determination of the anomalous magnetic momentof the muon to unprecedented precision. The precision of the g-2measurement will be increased by a factor of two over the mostrecent case (BNL, E821) mostly due to the increased statistics of-fered by the higher proton flux delivered by the Fermilab acceler-ators. However, it is possible that even further gains can be madethrough a better understanding of the muon beam being deliv-ered to the g-2 Storage Ring. Several effects come into play thatcan contribute to systematic errors and for which detailed calcu-lations and modeling of the incoming muon beam properties willaid in interpreting the results. Various correlations of spin andmomentum, spin and position along the bunch, etc., will becomeimportant to understand during the analysis of the experiment’sdata sets. While orders of magnitude of these types of effects are

44

straightforward to estimate, detailed calculations and experimen-tal verification of beam properties will be necessary to contributeto the sub-ppm accuracy of the g-2 measurement.

——————————————–

BEAM DYNAMICS SIMULATIONS OF THE MUONG-2 EXPERIMENT BEAMLINES AT FERMILAB

David Tarazona, Martin Berz, Kyoko Makino

Michigan State University

Classification: D-1, D-2, A-2

The main goal of the Muon g-2 Experiment (g-2) at Fermilab isto measure the muon anomalous magnetic moment to unprece-dented precision. This new measurement will allow to test thecompleteness of the Standard Model (SM) and to validate othertheoretical models beyond the SM. Simulations of the beamlinesfrom the pion production target to the entrance of the g-2 Stor-age Ring using COSY INFINITY have contributed to the under-standing of several factors that affect the systematics and statis-tical uncertainties of the anomaly measurement. Nonlinearities,spin dynamics, muon production, are some of the beam prop-erties analyzed from high-order simulations that consider fringefields, misalignments, and other errors intrinsic to the beamlines.

——————————————–

SEAMLESS BEAM AND RADIATION TRANSPORTSIMULATIONS OF IBA PROTEUS SYSTEMS

USING BDSIM

Robin Tesse (1), Stewart Boogert (2), Alain Dubus (1),Eustache Gnacadja (1), Cedric Hernalsteens (3), Laurence

J. Nevay (2), Nicolas Pauly (1), William Shields (2)

(1) Universite libre de Bruxelles (2) John Adams Instituteat Royal Holloway, University of London (3) Ion Beam

Applications (IBA)

Classification: A-2

The precise modeling of proton therapy systems is challengingand requires simulation tools that have capabilities in both beamtransport and in the detailed description of particle-matter inter-actions. Current separate simulations such as those of opticalcodes or Monte-Carlo transport through discrete elements showtheir limitations due to the very strict requirements on beam qual-ity at the isocenter. This is particularly relevant with the develop-ment of compact systems where the coupling between the com-ponents is dominant. For such systems the design of the concreteshielding, which has a large impact on the total cost of the sys-tem, is of primary importance. Beam Delivery Simulation (BD-SIM) allows the seamless simulation of the transport of particlesin a beamline and its surrounding environment. A complete 3Dmodel is built using Geant4, CLHEP and ROOT to provide anextensive insight into beam loss, its interaction and subsequentradiation. This capability is applied to the IBA eye treatmentproton therapy machine and to the IBA Proteus One compactsystem. We discuss the validation of both models against experi-mental data. In particular, we use it to predict lateral profiles andenergy spectra using a detailed geometry of the eye-treatment

beam forming nozzle. For the Proteus One system, we presentresults on the activation of the concrete shielding of the systemestimated after a period of 20 years of operation obtained for thefirst time using end-to-end simulations of the transport of protonsin the beamline and their interactions with the environment.

——————————————–

MULTI PASS ERL DESIGN WITH A SINGLEFIXED FIELD MAGNET RETURN LINE

Dejan Trbojevic, Stephen Brooks, Francois Meot, NickTsoupas, J. Scott Berg, William Lou(2)

BNL and Cornell University

Classification: A-2

We present a new approach of the Energy Recovery Linac De-sign for the future projects: PERLE (Powerful Energy Recov-ery Linac for Experiments), LHeC/FCCeH and eRHIC. The con-cept uses superconducting linacs and a single fixed field beamlines with multiple energy passes of electron beams. This rep-resents an update to the existing CBETA (Cornell UniversityBrookhaven National Laboratory ERL Test Accelerator) wherethe superconducting linac uses a single fixed field magnet beamline with four times in energy acceleration and four passes forthe energy recovery through the same single structure. To machthe single fixed field beam line to the linac the CBETA uses thespreaders and combiners on both sides of the linac, while the newconcept eliminates them. The arc cells from the single fixed fieldbeam line are connected to the linac with an adiabatic transitionarcs where its cells increase in length. The orbits of differentenergies merge into a single orbit through the interleaved linacwithin the straight sections as in the CBETA project. The beta-tron functions from the arcs are matched to the linac and the timeof flight of different electron energies is adjusted for the centralorbits by added kicker controlled induced beam oscillations.

——————————————–

COMPUTATION AND MEASUREMENT OFABERRATIONS FOR ABERRATION CORRECTED

ELECTRON MICROSCOPY

Rudolf M. Tromp

IBM T.J. Watson Research Center

Invited TalkClassification: A-2, E-2

Computation and measurement of geometric and chromatic aber-rations is critical for the optimal design and use of aberrationcorrected electron microscopes, and for quantitative understand-ing of images obtained with such instruments. Here, I will fo-cus on the correction of spherical and chromatic aberrations of acathode lens instrument (i.e. Low Energy Electron Microscope-LEEM- or Photo Electron Emission Microscope - PEEM) usingcatadioptrics, i.e. a combination of electron lenses (dioptrics)and an electron mirror (catoptrics). First-order properties cal-culated with high precision using Munro’s Electron Beam Soft-ware’s MIRDA package are in excellent with detailed experi-mental results. Theoretical maps of C3 vs Cc as a function of

45

the applied potentials then provide a deterministic method todial in the desired mirror properties at will. Now it is neces-sary to measure the resultant aberrations of the full system. Un-fortunately, the experimental methods developed for TEM andSTEM are not applicable in LEEM/PEEM for a variety of rea-sons. Spherical aberration (plus defocus and astigmatism) canbe measured using so-called micro-spot real-space Low EnergyElectron Diffraction, or by measuring image shift as a function ofbeam tilt. Measuring chromatic aberration is more troublesomeas it conventionally requires that defocus be measured as a func-tion of gun voltage. However, the use of magnetic prism arraysto separate in- and outgoing path in LEEM results in changingalignment conditions when gun voltage is changed. However,a novel method first demonstrated using ray-tracing simulationsenables us to measure chromatic aberration, even at fixed gunvoltage. The chromatically corrected system behaves like a sim-ple (but adjustable) achromat, comparable to the crown/flint op-tical achromat invented by Chester Moore Hall around 1730.

——————————————–

CALCULATION OF THE AGS OPTICS BASED ON3D FIELDS DERIVED FROM EXPERIMENTALLY

MEASURED FIELD MAPS ON THE MEDIANPLANE

N. Tsoupas†, J. S. Berg, S. Brooks, F. Meot, V. Ptitsyn, D.Trbojevic

Brookhaven National Laboratory

Classification: F-1

Closed orbit calculations of the AGS synchrotron were per-formed and the beam parameters at the extraction point of theAGS [1] were calculated using the RAYTRACE computer code[2] which was modified to generate 3D fields from the experi-mentally measured field maps on the median plane of the AGScombined function magnets. The algorithm which generates 3Dfields from field maps on a plane is described in reference [3]which discusses the details of the mathematical foundation ofthis approach. In this presentation we will discuss results fromstudies [1,4] that are based on the 3D fields generated from theknown field components on a rectangular grid of a plane. A briefoverview of the algorithm used will be given, and two methodsof calculating the required field derivatives on the plane will bepresented. The calculated 3D fields of a modified Halbach mag-net [5] of inner radius of 4.4 cm will be calculated using the twodifferent methods of calculating the field derivatives on the planeand the calculated fields will be compared against the “ideal”fields as calculated by the OPERA computer code [6]. *Worksupported by the US Department of Energy †tsoupas@bnl.gov[1] N. Tsoupas et. al. “Closed orbit calculations at AGS and Ex-traction Beam Parameters at H13” AD/RHIC/RD-75 Oct. 1994[2] S.B. Kowalski and H.A. Enge “The Ion-Optical ProgramRaytrace” NIM A258 (1987) 407 [3] K. Makino, M. Berz, C.Johnstone, Int. Journal of Modern Physics A 26 (2011) 1807-1821 [4] N. Tsoupas et. al. “Effects of Dipole Magnet Inhomo-geneity on the Beam Ellipsoid” NIM A258 (1987) 421-425 [5]“The CBETA project: arXiv.org > physics > arXiv:1706.04245”[6] Vector Fields Inc. https://operafea.com/

——————————————–

MAIN AND FRINGE FIELD COMPUTATIONS FORTHE ELECTROSTATIC QUADRUPOLES OF THE

MUON G-2 EXPERIMENT STORAGE RING

Eremey Valetov and Martin Berz

Michigan State University

Classification: C-2

We consider semi-infinite electrostatic deflectors with plates ofdifferent thickness, including plates with rounded edges, and wecalculate their electrostatic potential and field using conformalmappings. To validate the calculations, we compare the fringefields of these electrostatic deflectors with fringe fields of finiteelectrostatic capacitors, and we extend the study to fringe fieldsof adjacent electrostatic deflectors with consideration of electro-static induction, where field falloffs of semi-infinite electrostaticdeflectors are slower than exponential and thus behave differentlyfrom most magnetic fringe fields. Building on the success withelectrostatic deflectors, we develop a highly accurate and fullyMaxwellian conformal mappings method for calculation of mainfields of electrostatic particle optical elements. A remarkable ad-vantage of this method is the possibility of rapid recalculationswith geometric asymmetries and mispowered plates. We use thisconformal mappings method to calculate the multipole terms ofthe high voltage quadrupole used in the storage ring of the Muong-2 Experiment (FNAL-E-0989). Completing the methodolog-ical framework, we present a method for extracting multipolestrength falloffs of a particle optical element from a set of Fouriermode falloffs. We calculate the quadrupole strength falloff andits effective field boundary (EFB) for the Muon g-2 quadrupole,which has explained the experimentally measured tunes, whilesimple estimates based on a linear model exhibited discrepanciesup to 2%.

——————————————–

THE FAST/IOTA PROJECT AT FERMILAB

Alexander Valishev

Fermi National Accelerator Laboratory

Plenary TalkClassification: A-1

The Fermilab Accelerator Science and Technology (FAST) facil-ity is being developed as a fully-equipped accelerator chain in-tended to support research and development of accelerator tech-nology for the next generation of particle accelerators. The pri-mary focus of this effort is the Integrable Optics Test Accelerator(IOTA) ring, which will be able to circulate either electrons withthe energy of up to 150MeV, or 2.5MeV protons. The FAST elec-tron injector is a state of the art superconducting RF linac capableof full ILC beam parameters and beam energy of up to 300MeV.The FAST accelerator science program focuses on high-intensityand high-brightness issues in the future machines for high-energyphysics research. This talk will describe the facility design andstatus, review key beam physics experiments, and discuss thecomputational needs associated with the IOTA/FAST research.

——————————————–

46

PARALLEL ALGORITHMS FOR SOLVINGNONLINEAR EIGENVALUE PROBLEMS IN

ACCELERATOR CAVITY SIMULATIONS

Roel Van Beeumen

Lawrence Berkeley National Laboratory

Invited TalkClassification: F-2

We present an efficient and reliable algorithm for solving a classof nonlinear eigenvalue problems arising from the modeling ofparticle accelerator cavities. The eigenvalue nonlinearity in theseproblems results from the use of waveguides to couple externalpower sources or to allow certain excited electromagnetic modesto exit the cavity. We use a rational approximation to reduce thenonlinear eigenvalue problem first to a rational eigenvalue prob-lem. We then apply a special linearization procedure to turn therational eigenvalue problem into a larger linear eigenvalue prob-lem with the same eigenvalues, which can be solved by existingiterative methods. By using a compact scheme to represent boththe linearized operator and the eigenvectors to be computed, weobtain a numerical method that only involves solving linear sys-tems of equations of the same dimension as the original nonlin-ear eigenvalue problem. We refer to this method as a compactrational Krylov (CORK) method. We implemented the CORKmethod in the Omega3P module of the Advanced ComputationalElectromagnetic 3D Parallel (ACE3P) simulation suite and val-idated it by comparing the computed cavity resonant frequen-cies and damping Q factors of a small model problem to thoseobtained from a fitting procedure that uses frequency responsescomputed by another ACE3P module called S3P. We also usedthe CORK method to compute trapped modes damped in an idealeight 9-cell SRF cavity cryomodule. This was the first time it waspossible to compute these modes directly. The damping Q factorsof the computed modes match well with those measured in ex-periments and the difference in resonant frequencies is within therange introduced by cavity imperfection. Therefore, the CORKmethod is an extremely valuable tool for computational cavitydesign.

——————————————–

A NEW FINITE ELEMENT SOLVER FOR MOEVEPIC TRACKING

Ursula van Rienen1,2, Dawei Zheng1, Johann Heller1,Christian Bahls1

1 Institute of General Electrical Engineering, University ofRostock, D-18051 Rostock, Germany, 2 Department Life,Light & Matter, University of Rostock, 18051 Rostock,

Germany

Classification: D-1

A relevant task in designing high-brilliance light sources basedon high-current linear accelerators (e.g. Energy Recovery Linacs“ERLs”) consists in systematic investigations of ion dynamicsin the vacuum chamber of such machines. This is of high im-portance since the parasitic ions generated by the electron beamturned out to be a current-limiting factor for many synchrotron

radiation sources. In particular, the planned high current op-eration at ERL facilities requires a precise analysis and an ac-curate development of appropriate measures for the suppressionof ion-induced beam instabilities. The longitudinal transport ofions through the whole accelerator plays a key role for the es-tablishment of the ion concentration in the machine. Using theParticle-in-Cell (PIC) method, we redesigned our code MOEVEPIC Tracking in order to allow for the fast estimation of the ef-fects of ions on the beam dynamics. For that, we exchanged thepreviously used Finite Difference (FD) method for the solution ofPoisson’s equation within the PIC solver by a solver based on theFinite Element Method (FEM). Employing higher order FEM,we expect to gain improved convergence rates and thus lowercomputational times. We chose the Open Source FrameworkFEniCS for our new implementation. With regard to a better per-formance, we also studied an adaptive grid refinement togetherwith higher order approaches. We investigated certain strategieswith regard to compromises between accuracy and performancein the need of further refinement in a certain time step. Aiming toapply again the Boris pusher for the update in particle position,we examined several possibilities for an efficient determinationof the particle position.

——————————————–

BEAM DYNAMICS SIMULATIONS OF MEDICALCYCLOTRONS AND BEAM TRANSFER LINES AT

IBA

Jarno Van de Walle (1), Willem Kleeven (1), VincentNuttens (1), Erik Van der Kraaij (1) Jerome Mandrillon

(1), Eric Forton (1), Cedric Hernalsteens (1)

(1) Ion Beam Applications, Louvain-la-Neuve, Belgium

Classification: F-1, A-2, D-1

The company Ion Beam Applications (IBA), based in Belgium, isspecialized in the design and fabrication of cyclotrons for med-ical applications since more than 30 years. Two main classesof cyclotrons can be distinguished : cyclotrons for radiopharmaproduction (3 MeV up to 70 MeV proton beams) and cyclotronsused in proton therapy (230 MeV proton beam). In this con-tribution, the developments of computational tools to simulatebeam dynamics in the variety of cyclotrons and associated beamlines will be described. The main code for simulating the cy-clotron beam dynamics is the “Advanced Orbit Code” (AOC)[1]. Examples will be shown of beam dynamics studies in thenewly designed Cyclone KIUBE (18 MeV proton cyclotron forPET isotope production), the Cyclone230 and the superconduct-ing synchro-cyclotron (S2C2), both 230 MeV proton cyclotronsfor proton therapy. Calculated beam emittances, resonance cross-ings and beam losses will be shown and their impact on theperformance of the machine will be highlighted. A strong em-phasis will be put on the beam properties from the S2C2 (pro-ton therapy cyclotron), since unexpected extracted proton beamwas discovered and explained by detailed simulations [2] and thebeam properties serve as input to subsequent beam line simula-tion tools. Several tools have been developed to simulate anddesign transfer lines coupled to the cyclotrons. In radiopharmaapplications beam losses along the beamline and the beam sizeon the production target are crucial, since beam intensities are

47

high and radiation damage can be considerable. In proton ther-apy, beam intensities are very low but the constraints on the beamposition, drift (in position, energy and intensity) and size at thepatient level are very tight. In both cases a strong predictivepower of the calculated beam properties in the transfer lines isneeded. The compact proton gantry (CGTR) coupled with theS2C2 in the ProteusONE proton therapy system will be shownin detail. The CGTR is a spectrometer with sensitive beam di-agnostics devices and enables us to detect small fluctuations ofthe extracted beam properties. Measurements and calculations ofthe proton beam in the CGTR will be used to illustrate the per-formance of the calculation tools. [1] W. Kleeven et al., IPAC2016 proceedings, TUPOY002 [2] J. Van de Walle et al., Cy-clotrons2016 proceedings, THB01

——————————————–

MODE-ANALYSIS METHODS FOR THE STUDYOF COLLECTIVE INSTABILITIES IN

ELECTRON-STORAGE RINGS

Marco Venturini

LBNL

Classification: B-1, A-2, D-1

We report on recent progress on the application of mode analysisto the study of collective instabilities in electron storage rings in-cluding Higher Harmonic RF Cavities (HHCs). The focus is ontransverse instabilities in the presence of a dominant resistive-wall impedance, a problem of particular relevance to the newgeneration of diffraction-limited light sources. The secular equa-tion for the system eigenvalues is solved after applying a regu-larizing transformation, a key step to obtaining numerically ac-curate solutions. We provide a demonstration that for vanishingchromaticity and in the absence of radiation damping the beammotion is always unstable. This is in contrast to the more con-ventional Transverse-Mode-Coupling Instability (TMCI) with-out HHCs, which is known to exhibit a well defined instabilitythreshold.

——————————————–

ANALYTICAL CALCULATIONS FORTHOMSON-BACKSCATTERING BASED-LIGHT

SOURCES

Paul Volz, Atoosa Meseck

Helmholtz-Zentrum Berlin

Classification: F-1, A-2

There is a rising interest in Thomson-backscattering based-lightsources, as scattering intense laser radiation on MeV electronsproduces high energy photons that would require GeV or evenTeV electron beams when using conventional undulators ordipoles. Particularly, medium energy high brightness beamsdelivered by LINACs or Energy Recovery LINACs, such asbERLinPro being built at Helmholtz-Zentrum Berlin, seem suit-able for these sources. In order to study the merit of Thomson-backscattering-based light sources, we are developing an analyti-cal code to simulate the characteristics of the Thomson scattered

radiation. The code calculates the distribution of scattered ra-diation depending on the incident angle and polarization of thelaser radiation. Also the impact of the incident laser profile andthe full 6D bunch profile, including microbunching, are incorpo-rated. The Status of the code and first results will be presented.

——————————————–

A HOLISTIC APPROACH TO SIMULATING BEAMLOSSES IN THE LARGE HADRON COLLIDER

USING BDSIM

Stuart Walker, Andrey Abramov, Stewart Boogert,Hector Garcia Morales, Stephen Gibson, Helena

Pikhartova, William Shields, Laurie Nevay

Royal Holloway, University of London

Classification: F-1, A-2

To fully understand the beam losses, subsequent radiation, en-ergy deposition and activation in particle accelerators, a holisticapproach combining a 3-D model, physics processes and accel-erator tracking is required. Beam Delivery Simulation (BDSIM)is a program developed to simulate the passage of particles, bothprimary and secondary, in particle accelerators and calculate theenergy deposited by these particles via material interactions us-ing the Geant4 physics library. A Geant4 accelerator model isbuilt from an existing optical description of a lattice by procedu-rally placing a set of predefined accelerator components. Thesegeneric components can be refined to an arbitrary degree of detailwith the use of user-defined geometries, detectors, field maps,and more. A detailed model of the Large Hadron Collider hasbeen created in BDSIM, validated with existing tracking codesand applied to study beam loss patterns. The simulated beam lossmonitor response is compared with data from individual BLMsplaced around the LHC.

——————————————–

SHINE: SHANGHAI HIGH REP-RATE XFEL ANDEXTREME LIGHT FACILITY

Dong Wang and Weishi Wan

Chinese Academy of Sciences/ShanghaiTech University

Classification: A-1

SHINE(Shanghai High Rep-rate XFEL and Extreme Light Fa-cility) is a Free Electron Laser facility providing intense x-rayphotons at soft and hard X-ray regimes with high repetitionrate up to 1 MHz. This new facility is located at ZhangjiangNational Comprehensive Science Center, Shanghai, where alsohosts other large facilities on photon science including Shang-hai Synchrotron Radiation Facility(SSRF) and Soft X-ray FreeElectron Laser Facility(SXFEL). With an overall length of about3.1km the SHINE facility consists a linear accelerator yieldingup to 8 GeV electorn beam, 3 long FEL undulator lines produc-ing 0.4-25 keV coherent photons and 10 endstations for user ex-periments. The ground breaking of project took place in April,2018. This talk will present the status of SHINE facility with anemphasis on accelerator machine.

48

——————————————–

THEORETICAL AND COMPUTATIONALMODELING OF A PLASMA WAKEFIELD BBU

INSTABILITY

Stephen Webb, David Bruhwiler, Alexey Burov, andSergei Nagaitsev

RadiaSoft, LLC, RadiaSoft, LLC, Fermi NationalAccelerator Lab, Fermi National Accelerator Lab

Invited TalkClassification: B-2, D-1

Plasma wakefield accelerators achieve accelerating gradients onthe order of the wave-breaking limit, mc2kp/e, so that higher ac-celerating gradients correspond to shorter plasma wavelengths.Small-scale accelerating structures, such as plasma and dielec-tric wakefields, are susceptible to the beam break-up instabil-ity (BBU), which can be understood from the Panofsky-Wenzeltheorem: if the fundamental accelerating mode scales as b−1

for a structure radius b, then the dipole mode must scale asb−3, meaning that high accelerating gradients necessarily comewith strong dipole wake fields. Because of this relationship,any plasma-accelerator-based future collider will require detailedstudy of the trade-offs between extracting the maximum energyfrom the driver and mitigating the beam break-up instability.Recent theoretical work* predicts the tradeoff between the wit-ness bunch stability and the amount of energy that can be ex-tracted from the drive bunch, a so-called “efficiency-instabilityrelation”. We will discuss the beam break-up instability andthe efficiency-instability relation and the theoretical assumptionsmade in reaching this conclusion. We will also present prelimi-nary particle-in-cell simulations of a beam-driven plasma wake-field accelerator used to test the domain of validity for the as-sumptions made in this model. * V. Lebedev, A. Burov, and S.Nagaitsev, “Efficiency versus instability in plasma accelerators”,Phys. Rev. Acc. Beams 20, 121301 (2017).

——————————————–

ADVANCES IN ACCELERATOR MODELINGWITH PARALLEL MULTI-PHYSICS CODE SUITE

ACE3P

Liling Xiao, Lixin Ge, Zenghai Li, Cho-Kuen Ng

SLAC National Accelerator Laboratory

Classification: C-2, E-1

ACE3P is a comprehensive set of parallel finite-element codesfor multi-physics modeling of accelerator structures includingintegrated electromagnetic, thermal and mechanical effects. Re-cent advances of ACE3P have been focused on the developmentof multi-physics modeling capabilities, implementation of ad-vanced numerical algorithms, and improvement of code perfor-mance on state-of-the-art high-performance computing (HPC)platforms for large-scale accelerator applications. A nonlin-ear eigensolver using the CORK algorithm [1] has been imple-mented in the eigensolver module Omega3P to enable accuratedetermination of damping factors of resonant modes above the

beampipe cutoff frequency. It has enabled the first-ever directcalculation of trapped modes in the TESLA TTF cryomodules,providing reliable damping factors that were validated againstmeasurements. A newly developed mechanical eigensolver inthe multi-physics module TEM3P has allowed the determinationof mechanical modes in Fermilab PIP-II high beta 650 MHz cry-omodule, demonstrating mode coupling between the 6 cavities inthe cryomodule. To exploit multi-core computer architectures onsupercomputers, a hybrid MPI+OpenMP parallel programing hasbeen developed in the particle tracking module Track3P to speedup dark current simulation in multiple cavities for the LCLS-IIlinac. Highlights of these developments and their impacts onaccelerator modeling using HPC will be presented. [1] R. VanBeeuman, Invited talk, this conference.

——————————————–

MEAN-FIELD DENSITY EVOLUTION OFBUNCHED PARTICLES WITH NON-ZERO

INITIAL VELOCITY

Brandon Zerbe, Phil Duxbury

MSU

Classification: D-1

Reed(Reed 2006) presented a 1D mean-field model of initiallycold pancake-beam expansion demonstrating that the evolutionof the entire spatial distribution can be solved for all time wherethe 1D assumption holds. This model is relevant to ultra-fastelectron microscopy as it describes the evolution of the distri-bution within the photoelectron gun, and this model is similarto Anderson’s sheet beam density time dependence(Anderson1987) except that Reed’s theory applies to freely expandingbeams instead of beams within a focussing channel. Our re-cent work(Zerbe 2018) generalized Reed’s analysis to cylindricaland spherical geometries demonstrating the presence of a shockthat is seen in the Coulomb explosion literature under these ge-ometries and further discussed the absence of a shock in the 1Dmodel. This work is relevant as it offers a mechanistic explana-tion of the ring-like density shock that arises in non-equilibriumpancake-beams within the photoelectron gun; moreover, thisshock is coincident with a region of high-temperature electronsproviding an explanation for why experimentally aperturing theelectron bunch results in a greater than 10-fold improvementin beam emittance(Williams 2017), possibly even resulting inbunch emittance below the intrinsic emittance of the cathode.However, this theory has been developed for cold-bunches, i.e.bunches of electrons with 0 initial momentum. Here, we brieflyreview this new theory and extend the cylindrical- and spherical-symmetric distribution to ensembles that have non-zero initialmomentum distributions that are symmetric but otherwise unre-stricted demonstrating how initial velocity distributions couple tothe shocks seen in the less general formulation.

——————————————–

49

FAST MULTIPOLE METHODS FORMULTIPARTICLE SIMULATIONS

He Zhang

JLab

Invited TalkClassification: C-2, D-1, D-2

The fast multipole method (FMM) reduces the computation costof the pairwise non-oscillating interaction between N particlesfrom O(N2) to O(N). In the FMM, the contribution from a sourceparticle is represented as a multipole expansion, while the con-tributions from multiple faraway sources can be combined intoa local expansion around an objective particle. Without the de-pendence on a grid covering the whole domain under study, theFMM treats any charge distribution and geometry in a naturalway. It avoids artificial smoothing due to the grid size and redun-dant computation on the free space grids. We will introduce theconcept of the FMM using the Coulomb interaction as an exam-ple and then explain how the FMM can be extended to arbitrarynon-oscillating interactions. Examples and discussions on howthe FMM can be used in scientific simulations, especially in ac-celerator physics will also be provided.

——————————————–

GENERATION OF PARTICLE DISTRIBUTIONS ATRFQ EXIT AT SNS BEAM TEST FACILITY

Zhouli Zhang, A. Aleksandrov, S. Cousineau, A. Shishlo,A. Zhukov

University of Tennessee; Oak Ridge National Laboratory

Classification: A-2, E-1

The first 6D phase space measurement are being conducted atthe beam test facility (BTF) of SNS. Generation of 2D particledistributions are done first for preparation of reconstruction of6D distributions. A back-tracking PIC simulation code is writtenand proved to be reliable. The concept of distribution discrep-ancy is proposed to evaluate the effects of fluctuations of beamparameters and uncertainties of quadrupole gradients on initialdistributions at RFQ exit. Results suggest effects of fluctuationsof beam parameters are very small, while initial particle distri-butions are mainly affected by quadrupole gradients. The initialparticle distributions which are considered to be the closest to thereal ones are generated when distribution discrepancies are verysmall in transverse phase spaces and are proved to be convincingby comparing measured distributions and distributions producedby tracking the initial distributions. The distribution discrepancymethod in generation of initial particle distributions is confirmedto be practicable and can be used for reconstruction of 6D parti-cle distributions.

——————————————–

NUMERICAL SIMULATIONS FOR GENERATINGFULLY COHERENT SOFT X-RAY FREE

ELECTRON LASERS WITH ULTRA-SHORTWAVELENGTH

Kaishang Zhou

Shanghai Institute of Applied Physics, Chinese Academyof Sciences

Classification: B-1

For the fully coherent, ultra-short and high power soft x-rays arebecoming key instruments in different research fields, such as bi-ology, chemistry or physics. However it’s not easy to generatethis kind of advantaged light source by conventional lasers, espe-cially for the soft x-rays with ultra-short wavelength. The exter-nal seeded free electron laser (FEL) is considered as one feasiblemethod. Here, we give an example to generate fully coherentsoft x-rays with the wavelength 1nm by the two-stage cascadedFELs. The EEHG scheme is used in the first-stage while theHGHG scheme is used in the second-stage.

——————————————–

50

CPO-10 PARTICIPANTS

Abell, Dan (RadiaSoft)

Adelmann, Andreas (PSI)

Agemura, Toshihide (Hitachi)

Alimohamadi, Masoud (Farhangian University)

Asai, Hirotaka (Meijo University)

Balamuniappan, Pranesh (National University ofSingapore)

Barfels, Melanie (Gatan)

Berg, Georg (U of Notre Dame)

Berz, Martin (MSU)

Bimurzaev, Seitkerim (Almaty University of PowerEngineering and Telecommunication)

Carneiro, Jean-Paul (FNAL)

Cook, Nathan (RadiaSoft)

Dellby, Niklas (Nion)

Deng, Haixiao (SINAP)

Desai, Vishal (NuFlare Technology America, Inc.)

Edwards Jr, David (IJL Research Center)

Fedurin, Mikhail (BNL)

Friedman, Stuart (Nova Measuring Instruments, Inc.)

Fujita, Shin (Shimadzu)

Greenzweig, Yuval (Intel)

Grinfeld, Dmitry (Thermo Fisher Scientific)

Guo, Xiaoli (TU Delft)

Han, Wenjie (Huazhong University of Science andTechnology)

Herfurth, Frank (GSI)

Hesam Mahmoudi Nezhad, Neda (TU Delft)

Hoang, Hung Quang (Luxembourg Institute of Scienceand Technology)

Hoque, Shahedul (Hitachi)

Hornung, Christine (JLU Giessen)

Jiang, Hongping (Harbin Institute of Technology)

Jiang, Xinrong (KLA-TENCOR Corporation)

Jiao, Yi (Institute of High Energy Physics, CAS)

Jiruse, Jaroslav (TESCAN)

Jung, Paul (TRIUMF)

Kahl, Frank (CEOS)

Katsap, Victor (NuFlare Technology America)

Kazantseva, Erika (TU Darmstadt)

Khan, Sameen (Dhofar University)

Khedher Agha, Ahmad (Al-Nahrain University)

Khursheed, Anjam (National University of Singapore)

Kim, Jong-Won (Institute for Basic Science)

Kozak, Martin (Charles University)

Krivanek, Ondrej (Nion)

Kruit, Pieter (TU Delft)

Lebedev, Gennadi (Getom Analytical Instrumentations)

Liu, Bo (SINAP)

Liu, Xu (Huazhong University of Science andTechnology)

Liu, Xuedong (Hermes Microvision Inc.)

Lombardi, Alexander (ZEISS)

Luiten, Jom (TU Eindhoven)

Lunin, Andrei (FNAL)

Lyman, Charles (Microscopy Society of America)

Maazouz, Mostafa (Thermo Fisher Scientific)

Makino, Kyoko (MSU)

Manikonda, Shashikant (Magtech Solutions LLC)

Mankos, Marian (Electron Optica, Inc.)

Maxson, Jared (Cornell U)

Mayes, Christopher (SLAC)

McClelland, Jabez (NIST)

McGinn, Jim (Thermo Fisher Scientific)

Munro, Eric (Munro’s Electron Beam Software)

Murata, Hidekazu (Meijo University)

Nesteruk, Konrad (PSI)

Neustock, Lars Thorben (Stanford U)

Nisenboim, Evgeny (Intel)

Nishi, Ryuji (Osaka University)

Ogasawara, Munehiro (NuFlare Technology)

51

Ogawa, Takashi (Korea Research Institute of Standardsand Science)

Oral, Martin (Institute of Scientific Instruments, CAS)

Ose, Yoichi (Hitachi)

Qin, Bin (Huazhong University of Science andTechnology)

Radlicka, Tomas (Institute of Scientific Instruments, CAS)

Ren, Weiming (Hermes Microvision Inc.)

Ren, Yan (ASML)

Rosenbusch, Marco (RIKEN)

Rosenthal, Marcel (CERN)

Rouse, Catherine (Munro’s Electron Beam Software)

Rouse, John (Munro’s Electron Beam Software)

Russell, Zach (Ion Innovations)

Seda, Bohuslav (Thermo Fisher Scientific)

Seidling, Michael (Friedrich-Alexander UniversityErlangen-Nuremberg)

Shadman, Khashayar (Electron Optica)

Shchepunov, Vyacheslav (Shimadzu)

Shirasaki, Yasuhiro (Hitachi)

Spivak-Lavrov, Igor (Aktobe State University)

Stewart, Hamish (Thermo Fisher Scientific)

Stopka, Jan (Institute of Scientific Instruments, CAS)

Tamaki, Hirokazu (Hitachi)

Tarazona, David (MSU)

Tesch, Paul (Oregon Physics)

Trbojevic, Dejan (BNL)

Tromp, Rudolf (IBM)

Tsoupas, Nicholaos (BNL)

Turchetti, Marco (MIT)

Valetov, Eremey (MSU)

Van De Walle, Jarno (Ion Beam Applications)

Vasina, Radovan (Thermo Fisher Scientific)

Vering, Guido (Raith GmbH)

Wan, Weishi (ShanghaiTech University)

Watts, Adam (FNAL)

Webb, Stephen (RadiaSoft)

Weel, Matthew (Fluidigm Canada)

Weisskopf, Adrian (MSU)

Wieland, Marco (Mapper Lithography BV)

Wollnik, Hermann (New Mexico State U)

Yamazawa, Yu (Hitachi)

Zerbe, Brandon (MSU)

Zhang, Zhouli (U of Tennessee)

Zhou, Kaishang (SINAP)

52

ICAP’2018 PARTICIPANTS

Abell, Dan (RadiaSoft)

Adelmann, Andreas (PSI)

Aleksandrov, Alexander (ORNL)

Alimohamadi, Masoud (Farhangian University)

Appel, Sabrina (GSI)

Bassi, Gabriele (BNL)

Berz, Martin (MSU)

Beznosov, Oleksii (U of New Mexico)

Bizzozero, David (TU Darmstadt)

Boine-Frankenheim, Oliver (TU Darmstadt)

Buescher, Markus (FZ Julich)

Cerfon, Antoine (New York U)

Cook, Nathan (RadiaSoft)

De Gersem, Herbert (TU Darmstadt)

De Maria, Riccardo (CERN)

Deng, Haixiao (SINAP)

Deniau, Laurent (CERN)

Edelen, Auralee (SLAC)

Ellison, Jim (U of New Mexico)

Erdelyi, Bela (NIU)

Fedurin, Mikhail (BNL)

Flisgen, Thomas (Helmholtz-Zentrum Berlin)

Frey, Matthias (PSI)

Friedman, Stuart (Nova Measuring Instruments, Inc.)

Gjonaj, Erion (TU Darmstadt)

Gulliford, Colwyn (Cornell U)

Han, Baoxi (ORNL)

Han, Wenjie (Huazhong University of Science andTechnology)

Hassan, Amin (Egyptian Atomic Energy Authority)

Heinemann, Klaus (U of New Mexico)

Hernalsteens, Cedric (IBA)

Hidas, Dean (BNL)

Huggins, Anthony (HHU Dusseldorf)

Hwang, Kilean (LBNL)

Iwasaki, Yoshitaka (SAGA Light Source)

Jensen, Aaron (Leidos)

Jiang, Hongping (Harbin Institute of Technology)

Johnstone, Carol (FNAL)

Jung, Paul (TRIUMF)

Kazantseva, Erika (TU Darmstadt)

Khedher Agha, Ahmad (Al-Nahrain University)

Kramer, Patrick (CERN)

Kranjcevic, Marija (ETH Zurich)

Krasilnikov, Mikhail (DESY)

Kunz, Josiah (Anderson U)

Lehe, Remi (LBNL)

Levinsen, Yngve (European Spallation Source)

Li, Yongjun (BNL)

Liu, Ao (Euclid Techlabs)

Liu, Bo (SINAP)

Liu, Xu (Huazhong University of Science andTechnology)

Lunin, Andrei (FNAL)

Luo, Yun (BNL)

Ma, Jun (BNL)

Maazouz, Mostafa (Thermo Fisher Scientific)

Makino, Kyoko (MSU)

Manikonda, Shashikant (Magtech Solutions LLC)

Marocchino, Alberto (INFN)

Massimo, Francesco (LLR - CNRS)

Mayes, Christopher (SLAC)

Meot, Francois (BNL)

Metral, Elias (CERN)

Mitchell, Chad (LBNL)

Munro, Eric (Munro’s Electron Beam Software)

Nash, Boaz (RadiaSoft)

Neveu, Nicole (IIT)

Niedermayer, Uwe (TU Darmstadt)

Orozco, Eduardo (Universidad Industrial de Santander)

53

Ostroumov, Peter (MSU)

Otero Olarte, Oswaldo (Universidad Industrial deSantander)

Park, Chong Shik (Korea University)

Planche, Thomas (TRIUMF)

Plastun, Alexander (FRIB/MSU)

Pogorelov, Ilya (RadiaSoft)

Pommerenke, Hermann (CERN, University of Rostock)

Qiang, Ji (LBNL)

Qin, Bin (Huazhong University of Science andTechnology)

Ranjbar, Vahid (BNL)

Ratner, Daniel (SLAC)

Repond, Joel (CERN)

Rosenthal, Marcel (CERN)

Rouse, Catherine (Munro’s Electron Beam Software)

Rouse, John (Munro’s Electron Beam Software)

Ruisard, Kiersten (ORNL)

Russenschuck, Stephan (CERN)

Ryne, Robert (LBNL)

Sagan, David (Cornell University)

Schmid, Steffen (TU Darmstadt)

Shishlo, Andrei (ORNL)

Snopok, Pavel (IIT)

Syphers, Michael (NIU)

Tarazona, David (MSU)

Tesse, Robin (Universite Libre de Bruxelles)

Trbojevic, Dejan (BNL)

Tromp, Rudolf (IBM)

Tsoupas, Nicholaos (BNL)

Valetov, Eremey (MSU)

Valishev, Alexander (FNAL)

Van Beeumen, Roel (LBNL)

Van De Walle, Jarno (Ion Beam Applications)

Van Rienen, Ursula (University of Rostock)

Venturini, Marco (LBNL)

Volz, Paul (Helmholtz-Zentrum Berlin)

Walker, Stuart (Royal Holloway, University of London)

Wan, Weishi (ShanghaiTech University)

Wang, Dong (Chinese Academy of Sciences,ShanghaiTech University)

Wang, Xiaofei (China Institute of Atomic Energy)

Webb, Stephen (RadiaSoft)

Weisskopf, Adrian (MSU)

Xiao, Liling (SLAC)

Zerbe, Brandon (MSU)

Zhang, He (Jlab)

Zhang, Zhouli (U of Tennessee)

Zhou, Kaishang (SINAP)

54

Index

Abedzadeh, N., 20Abell, D. T., 23, 41Abramov, A., 48Acharya, D., 28Ackermann, W., 27Adelmann, A., 23, 28–30, 34Agarwal, A., 20Aldiyarov, N. U., 1Aleksandrov, A., 23, 50Alimohamadi, M., 24Andersson, J., 27Andre, K., 31Ang, W. K., 9Appel, S., 24Appelo, D., 24, 31Arbenz, P., 34Arsenyev, S., 25Asai, H., 1, 13Astapovych, D., 25Audinot, J.-N., 5

Baartman, R., 33Bahls, C., 47Banerjee, D., 43Barber, D., 24, 31Bartnik, A., 30Bassi, G., 24Beaudoin, B., 43Beck, A., 37Beekman, I., 23Beltran, J. R., 40Berg, G. P. A., 1Berg, J. S., 30, 45, 46Berggren, K. K., 20Berglyd Olsen, V. K., 27Bernhard, J., 43Berz, M., 20, 21, 33, 45, 46Beznosov, O., 24, 31Bimurzaev, S. B., 1, 2Bizzozero, D., 25, 30Blaskiewicz, M., 20Bleloch, A. L., 9Boine-Frankenheim, O., 8, 25Boogert, S., 45, 48Boonpornprasert, P., 34Brooks, S., 20, 38, 45, 46Brouwer, L., 32Brugger, M., 43Bruhwiler, D., 41, 49Buescher, M., 26Burkhardt, H., 28Burov, A., 49

Caiafa, G., 43Calzolaio, C., 13

Carlsson, J., 41Carneiro, J.-P., 2Cathey, B., 23Cauchois, A., 37Cerfon, A., 26Charitonidis, N., 43Chen, Q., 35Chen, S., 35Cheng, W., 35Chetvertkova, V., 25Chung, M., 41Collignon, V., 31Cook, N., 26Corno, J., 27Cousineau, S., 23, 50Crittenden, J., 38

De Gersem, H., 25, 27, 30, 44De Maria, R., 27De Prisco, R., 34Dellby, N., 9Deng, H., 28, 35Deniau, L., 28Denisov, E., 4Derouillat, J., 37Dickel, T., 6Dobbins, J., 30Dohi, H., 10Drezner, Y., 3Dubus, A., 45Duxbury, P., 49Dobrich, B., 43

Eckstein, T., 9Edelen, A., 28Edelen, J., 26Edwards, D., Jr., 2Eidelman, Y., 41Ellison, J. A., 24, 31Enyama, M., 18Erdelyi, B., 29Eshraqi, M., 34Eswara, S., 5Ezumi, M., 15

Fedurin, M., 3, 29Feng, C., 35Feng, G., 35Field, L., 27Flandroy, Q., 31Flisgen, T., 29Forton, E., 47Freemire, B., 41Frey, M., 28–30Fujita, R., 17Fujita, S., 3

Fuoss, P., 37

Gardner, J. R., 12Gatignon, L., 43Ge, L., 49Geithner, W., 24Georg, N., 27Gerbershagen, A., 13, 43Gerity, J., 41Ghaffarian Niasar, M., 5Gibson, S., 48Giles, R., 17Giovannozzi, M., 27, 28Gjonaj, E., 25, 30, 44Gnacadja, E., 45Gonzalez, J. D., 40Gorgi Zadeh, S., 27, 34Grandsaert, T., 34Grech, M., 37Greenzweig, Y., 3Grinfeld, D., 4, 19Gu, Q., 35Gulliford, C., 30Guo, X., 4

Haber, I., 43Hagen, C.W., 5Hall, C., 26, 41Hallstein, R. M., 3Han, B., 30Han, W., 31Hanna, B., 2Hansen, P. C., 14Hao, Y., 36Heinemann, K., 24, 31Heller, J. D., 42, 47Herfurth, F., 5Hermes, P. D., 27Hernalsteens, C., 31, 45, 47Herrengods, B., 31Herrera, A. M., 40Hesam Mahmoudi Nezhad, N., 5Hesselink, L., 14Hidas, D. A., 32Hirayama, Y., 19Hoang, H. Q., 5Hock, C., 19Hoffman, M. V., 9Hommelhoff, P., 9Hoque, S., 6, 14Hornung, C., 6Huang, N., 28Hudson, S., 39Huetzen, A., 26Huggins, A., 32

55

Hwang, K., 32Høimyr, N., 27

Iadarola, G., 27Ikeda, T., 13Iliakis, K., 42Inada, H., 19Ishizawa, S., 22Ito, H., 6, 14Ito, Y., 16, 22Ives, L., 33Iwasaki, Y., 32

Jagannathan, R., 8Janssen, T., 5Jansson, A., 34Jensen, A., 33Jiao, Y., 7Jing, C., 35Jiruse, J., 7Johnstone, C., 21, 33Jowett, J. M., 28Juffmann, T., 12Jung, P. M., 33Jungers, R., 31

Kahl, F., 7Karpov, I., 42Kasevich, M. A., 12Katsap, V., 8Kawai, R., 1Kazantseva, E., 8Kean, A. W., 15Keilman, M., 26Khabiboulline, T., 36Khan, S. A., 8Khojoyan, M., 37Khursheed, A., 9, 15Kim, J., 9Kimura, S., 22Kleeven, W., 47Klopfer, B. B., 12Knuffman, B., 12Kocevar, H., 34Koeth, T., 43Kolarik, V., 12Koppell, S. A., 12Kornilov, V., 25Kostoglou, S., 27Kozak, M., 9Kramer, P., 33Kranjcevic, M., 34Krasinikov, M., 34Krielaart, M., 10Krivanek, O., 9Kruit, P., 4, 5, 10, 11, 19

Larson, J., 39

Lasheen, A., 42Latina, A., 28Laur, D., 11Lehe, R., 34Lehrach, A., 26Levinsen, Y., 34Li, K., 28Li, W., 12Li, Y., 35Li, Z., 49Liang, Z., 31Liebsch, M., 43Linck, M., 7Litvinenko, V. N., 36Liu, A., 35Liu, B., 35Liu, H., 13Liu, K., 31Liu, X., 31, 35Livengood, R. H., 3Lou, W., 20, 45Lovejoy, T. C., 9Lovelace, H., III, 20Luiten, J., 11Lunin, A., 36Luo, Y., 36Ly, Minh P., 3Lyman, C. E., 11

Ma, J., 36Maclean, E. H., 27Makarov, A., 4, 19Makino, K., 20, 33, 45Mandrillon, J., 47Mankos, M., 12, 17Marocchino, A., 36Massimo, F., 37Matsubara, F., 1Maxson, J., 12Mayes, C., 37McClelland, J., 12McGehee, W. R., 12Mcintosh, E., 27McIntyre, P., 41Meer, D., 13Meot, F., 20, 23, 38, 45, 46Mereghetti, A., 27Meseck, A., 48Metral, E., 38Midttun, O., 34Milas, N., 34Mitchell, C., 39, 41Miyamoto, R., 34Miyatake, H., 22Moeller, P., 26Mohammadi-Gheidari, A., 4, 5Molson, J., 27

Montag, C., 20Montbarbon, E., 43Morales, H. G., 48Munro, E., 13Murata, H., 1, 13Murray, S. N., Jr., 30Mustapha, B., 9Mutsaers, P., 11Muller, H., 7

Nagaitsev, S., 49Nagler, R., 23, 26Nakanishi, H., 17Nash, B., 39Neilson, J., 33Nesteruk, K. P., 13Neustock, L. T., 14Nevay, L. J., 45, 48Neveu, N., 28, 39Ng, C.-K., 49Niedermayer, U., 25, 39Nishi, R., 6, 14Nuttens, V., 47

Ogawa, T., 14Oral, M., 15Orozco, E. A., 40Ose, Y., 15Ostroumov, P. N., 40, 41Otero Olarte, O., 40

Park, C. S., 41Pauly, N., 45Pellegrini, D., 27Penn, G., 41Pennisi, T. R., 30Peplov, V., 30Persson, T., 27, 28Petillo, J., 33Petrone, C., 43Pham-Xuan, V., 27Pikhartova, H., 48Planche, T., 33, 41Plastun, A. S., 41Plostinar, C., 34Pogorelov, I., 41Pommerenke, H. W., 42Ponton, A., 34Pranesh, B., 15Price, R. L., 11Prost, L., 2Prudent, J., 37Ptitsyn, V., 20, 46Perez, F., 37

Qiang, J., 32, 36, 39, 42Qin, B., 16, 31, 35Qin, Q., 7

56

Quartulo, D., 42

Radlicka, T., 16Rae, B., 43Rainer, R., 35Ranjbar, V., 20, 42Ratner, D., 42Raveh, A., 3Read, M., 33Reimann, S., 24Repond, J., 42Rignall, M., 17Rizzoglio, V., 13Roblin, Y., 36Rokuta, E., 1, 13Rosenbusch, M., 16, 22Rosenthal, M., 43Rouse, C., 13Rouse, J., 13Rouson, D., 23Ruisard, K., 43Russell, Z. E., 14Russenschuck, S., 43Ryne, R. D., 44Romer, U., 27

Saethre, R., 30Sagan, D., 30Saini, A., 2Sapargaliev, A. A., 18Sapinski, M., 24Schillinger, R., 7Schippers, J. M., 13Schmid, S., 30, 44Schmidt, F., 28Schulte, D., 25Schury, P., 16, 22Schwarz, M., 42Schwarzkopf, A., 12Schwinzerl, M., 27Schonenberger, N., 9Schops, S., 27Seidel, M., 13Shadman, K., 12, 17Shao, X., 9Shaposhnikova, E., 42Shchepunov, V., 17Shea, T., 34Shemyakin, A., 2Shields, W., 45, 48Shimoyama, H., 13

Shirasaki, Y., 18Shishlo, A., 44, 50Singh, R., 24Singh, S., 27Sjobak, K., 27Skoblin, M., 4Skowronski, P. K., 28Snopok, P., 33Snuverink, J., 29Solyak, N., 36Sorge, S., 25Spadaro, T., 43Specka, A., 37Spentzouris, L., 39Spivak-Lavrov, I. F., 18Stancari, G., 41Steele, A. V., 12Stern, E., 41Stewart, H., 4, 19Stinson, C. M., 30Stockli, M. P., 30Stopka, J., 19Su, C., 9Sukhanov, A., 36Sulimov, A., 27Sun, D., 2Syphers, M., 44

Takaoka, A., 6, 14Tamaki, H., 19Tan, S., 3Tarazona, D., 20, 21, 45Tepikian, S., 20Tesse, R., 31, 45Thomas, J., 26Thomsen, H. D., 34Timko, H., 42Trbojevic, D., 20, 38, 45, 46Tromp, R. M., 45Trubitsyn, A. A., 18Tsoupas, N., 20, 38, 45, 46Turchetti, M., 20

Valetov, E., 21, 46Valishev, A., 46Van Beeumen, R., 47Van de Walle, J., 47Van der Kraaij, E., 47Van Dijk, M., 43Van Kouwen, L., 11Van Rens, J., 11

Van Rienen, U., 27, 34, 42, 47Venturini, M., 48Vergara, V. E., 40Verhoeven, W., 11Villari, A. C. C., 41Vilsmeier, D., 24Vollinger, C., 33Volz, P., 48

Wada, M., 16, 22Waki, H., 17Walker, S., 48Wan, W., 32, 48Wang, D., 35, 48Wang, E., 20Wang, G., 20, 36Wang, Z., 31Wang, Z., 35Watanabe, S., 13Watts, A., 21Webb, S. D., 21, 41, 49Weick, H., 8Weisskopf, A., 21Welton, R. F., 30Weng, W., 20Wieland, M., 22Willeke, F., 20Wirtz, T., 5Wollnik, H., 16, 22

Xiao, L., 49Xu, G., 7

Yakovlev, V., 36Yakushev, E. M., 1, 18Yang, J., 31Yaoshuo, Y., 25Yoon, C., 37Yu, L. H., 35

Zacharov, I., 27Zemzemi, I., 37Zerbe, B., 49Zhang, H., 50Zhang, M., 35Zhang, W., 20Zhang, Z., 23, 50Zhao, Q., 41Zhao, Z., 35Zheng, D., 47Zhou, K., 22, 50Zhukov, A., 23, 50

57

Contents

CPO-10 Abstracts 1

ICAP’2018 Abstracts 23

CPO-10 Participants 51

ICAP’2018 Participants 53